diff --git a/README.md b/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..a1701ef3962c9db58a41bf5ae6ab54bb75ac7f8f
--- /dev/null
+++ b/README.md
@@ -0,0 +1,104 @@
+---
+license: apache-2.0
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-classification
+language:
+  - en
+tags:
+  - decision-model
+  - system-one
+  - lora
+  - calibrated
+  - classification
+  - routing
+---
+
+# Certus v0007 — a "System One" decision model
+
+Try it: **[ait-hf/certus-playground](https://huggingface.co/spaces/ait-hf/certus-playground)** (free, no account).
+
+Certus is a self-hosted decision model in the spirit of TypeSafe's Jev / System One models. It does **not generate
+text**: it reads a `state` (any text or JSON) plus typed questions and returns **calibrated probability
+distributions** over answer sets the caller defines — a choice among named options, a score on an ordered rubric,
+or a yes/no. Every answer is the soft-max over the logits of the option letters at the assistant turn, so the output
+is always one of your options and the cost is one forward pass, no decoding.
+
+## What is in this repo
+
+```
+registry.json                 version registry (lineage, tasks, calibration temperatures, metrics)
+v0007/                        trunk: LoRA r=16 adapter on Qwen2.5-1.5B-Instruct, 61 public tasks, 1 epoch
+v0007-<domain>/               domain adapters (LoRA, parents = [v0007]); applied unmerged on top of the merged trunk
+family_v0007/family.json      domain -> adapter map, routing threshold, max stack
+family_v0007/heads/           router: one logistic head per adapter on the trunk's last hidden state + normalisation
+```
+
+| version | domain | trained on | LoRA r | val acc (mean over its tasks) |
+|---|---|---|---:|---:|
+| `v0007` | trunk | ag_news, banking77, boolq, clinc150, commonsense_qa, copa, dbpedia, emotion, facts, fits, imdb, jailbreak, massive_intent, mnli, mrpc, openbookqa, paws, qnli, read, rte, sciq, sst2, swag, tweet_emoji, tweet_hate, tweet_irony, tweet_offensive, tweet_sentiment, yahoo, yelp, anli, winogrande, hellaswag, race, scitail, qqp, stsb, toxic, stance_abortion, stance_atheism, stance_feminist, stance_hillary, match, goemo_soft, reason, formality, politeness, strategyqa, vitaminc, ruletaker, proofwriter, folio, logiqa, tracie, temporal_nli, piqa, siqa, clutrr, gsm8k, svamp, aqua | 16 | 0.792 |
+| `v0007-reason` | reason | traps, reason, skills | 16 | 0.943 |
+| `v0007-math` | math | math, gsm8k, svamp, aqua | 32 | 0.690 |
+| `v0007-language` | language | language | 8 | 1.000 |
+| `v0007-support` | support | banking77, clinc150, massive_intent | 8 | – |
+| `v0007-sentiment` | sentiment | sst2, sst5, yelp, tweet_sentiment, emotion, goemo_soft, tweet_irony, imdb, formality, politeness, sarcasm | 8 | 0.776 |
+| `v0007-safety` | safety | jailbreak, toxic, tweet_hate, tweet_offensive | 8 | 0.844 |
+| `v0007-spam` | spam | spam | 8 | 0.990 |
+| `v0007-typed` | workflow | typed_decisions | 16 | 0.850 |
+
+Serving: the trunk adapter is merged into the base model (base speed); the domain adapters stay separate and are
+switched per request. `model: "auto"` runs the router on the **full request** (every question with its type and
+options, then the state), applies every adapter whose head fires above 0.5 (at most two, stacked), and answers with
+the trunk when none does. Adding an adapter later means adding one folder and one head file — no retraining.
+
+## Results (v0007, hand-written held-out sets, never trained on)
+
+| set | trunk `v0007` | `auto` (router + adapters) |
+|---|---:|---:|
+| 196-question test bench, 20 categories | 0.842 | **0.949** |
+| typed decisions (Laya benchmark, 1 200-case format) | – | 0.796 (Jev 0.727, Laya 0.766) |
+| 10 hard yes/no logic traps (noul10) | 4/10 | 9/10 (Jev 10/10) |
+| 15-task unseen public suite, mean acc / ECE | 0.736 / 0.074 | – |
+
+Calibration: temperature scaling per option-count bucket (`config.json`); the trunk's ECE on its validation set is
+0.010 before scaling.
+
+## Use
+
+The inference code lives in the Space (`s1/` package, plain Python on top of `transformers` + `peft`).
+Minimal reproduction of the scoring rule without it:
+
+```python
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from peft import PeftModel
+
+base = "Qwen/Qwen2.5-1.5B-Instruct"
+tok = AutoTokenizer.from_pretrained(base)
+model = PeftModel.from_pretrained(AutoModelForCausalLM.from_pretrained(base, dtype=torch.float16), "v0007/adapter").merge_and_unload().cuda().eval()
+
+prompt = tok.apply_chat_template([
+    {"role": "user", "content": "Text:\nHelp! My payouts have been failing for 3 days.\n\nWhich team should handle this?\nA. billing\nB. technical\nC. sales\nAnswer with the letter."}],
+    tokenize=False, add_generation_prompt=True)
+ids = tok(prompt, return_tensors="pt").to("cuda")
+logits = model(**ids).logits[0, -1]
+letters = [tok.encode(l, add_special_tokens=False)[0] for l in "ABC"]
+print(torch.softmax(logits[letters] / 1.02, -1))          # 1.02 = calibration temperature for 3-5 options
+```
+
+The exact prompt layout (state rendering, option descriptions, yes/no order, the two-stage path for > 26 options)
+is in the Space's `s1/lmscore.py` and `s1/family.py`.
+
+## Training data
+
+Trunk: 61 public datasets (sentiment, topic, NLI, boolean QA, intents, commonsense/science QA, logic and word
+problems, safety) at ≤ 3 000 examples each, plus synthetic reasoning probes, with LoRA r=16, a proper-scoring
+(NLL + Brier) loss, yes/no balancing and a KL anchor to the base model. Adapters: see the table above. No benchmark
+item in the results section was used for training. Datasets with non-commercial licences were kept out of this release.
+
+## Limits
+
+- A 1.5B single-pass reader: pure computation (prices × quantities, unit conversions) and multi-step logic traps are
+  hit-and-miss even with the math and reasoning adapters.
+- States are truncated to 700 tokens.
+- English first; a multilingual adapter is planned for the next version.
diff --git a/family_v0007/family.json b/family_v0007/family.json
new file mode 100644
index 0000000000000000000000000000000000000000..c191e53235f3f11c29b95153bde21d7202414528
--- /dev/null
+++ b/family_v0007/family.json
@@ -0,0 +1,55 @@
+{
+  "trunk": "v0007",
+  "domains": {
+    "reason": "v0007-reason",
+    "math": "v0007-math",
+    "language": "v0007-language",
+    "support": "v0007-support",
+    "sentiment": "v0007-sentiment",
+    "safety": "v0007-safety",
+    "spam": "v0007-spam",
+    "workflow": "v0007-typed"
+  },
+  "route_threshold": 0.5,
+  "router": {
+    "accuracy": {
+      "train": 0.9978819444444444,
+      "val": 0.9866666666666667,
+      "test": 0.9883333333333333
+    },
+    "confusion": {
+      "sentiment": {
+        "sentiment": 74,
+        "reason": 1
+      },
+      "support": {
+        "support": 75
+      },
+      "reason": {
+        "reason": 73,
+        "spam": 1,
+        "general": 1
+      },
+      "finance": {
+        "finance": 74,
+        "support": 1
+      },
+      "language": {
+        "language": 75
+      },
+      "general": {
+        "general": 73,
+        "reason": 1,
+        "sentiment": 1
+      },
+      "safety": {
+        "safety": 74,
+        "reason": 1
+      },
+      "spam": {
+        "spam": 75
+      }
+    }
+  },
+  "created": "2026-09-22 09:56"
+}
\ No newline at end of file
diff --git a/family_v0007/router_head.pt b/family_v0007/router_head.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b9d675f56b57efabb0351cc57ca71275b19c21a1
--- /dev/null
+++ b/family_v0007/router_head.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbabf53a1780e64af4408ecbadb3c1289f4bc7358f7881050fdbe7546b846d67
+size 64837
diff --git a/registry.json b/registry.json
new file mode 100644
index 0000000000000000000000000000000000000000..8d21ea1928a3174ed9c327485664cdbe8887ab41
--- /dev/null
+++ b/registry.json
@@ -0,0 +1,2251 @@
+{
+  "latest": "v0007",
+  "versions": {
+    "v0007": {
+      "version": "v0007",
+      "parent": null,
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-17T15:58:07+00:00",
+      "tasks": [
+        "ag_news",
+        "banking77",
+        "boolq",
+        "clinc150",
+        "commonsense_qa",
+        "copa",
+        "dbpedia",
+        "emotion",
+        "facts",
+        "fits",
+        "imdb",
+        "jailbreak",
+        "massive_intent",
+        "mnli",
+        "mrpc",
+        "openbookqa",
+        "paws",
+        "qnli",
+        "read",
+        "rte",
+        "sciq",
+        "sst2",
+        "swag",
+        "tweet_emoji",
+        "tweet_hate",
+        "tweet_irony",
+        "tweet_offensive",
+        "tweet_sentiment",
+        "yahoo",
+        "yelp",
+        "anli",
+        "winogrande",
+        "hellaswag",
+        "race",
+        "scitail",
+        "qqp",
+        "stsb",
+        "toxic",
+        "stance_abortion",
+        "stance_atheism",
+        "stance_feminist",
+        "stance_hillary",
+        "match",
+        "goemo_soft",
+        "reason",
+        "formality",
+        "politeness",
+        "strategyqa",
+        "vitaminc",
+        "ruletaker",
+        "proofwriter",
+        "folio",
+        "logiqa",
+        "tracie",
+        "temporal_nli",
+        "piqa",
+        "siqa",
+        "clutrr",
+        "gsm8k",
+        "svamp",
+        "aqua"
+      ],
+      "trained_on": [
+        "ag_news",
+        "banking77",
+        "boolq",
+        "clinc150",
+        "commonsense_qa",
+        "copa",
+        "dbpedia",
+        "emotion",
+        "facts",
+        "fits",
+        "imdb",
+        "jailbreak",
+        "massive_intent",
+        "mnli",
+        "mrpc",
+        "openbookqa",
+        "paws",
+        "qnli",
+        "read",
+        "rte",
+        "sciq",
+        "sst2",
+        "swag",
+        "tweet_emoji",
+        "tweet_hate",
+        "tweet_irony",
+        "tweet_offensive",
+        "tweet_sentiment",
+        "yahoo",
+        "yelp",
+        "anli",
+        "winogrande",
+        "hellaswag",
+        "race",
+        "scitail",
+        "qqp",
+        "stsb",
+        "toxic",
+        "stance_abortion",
+        "stance_atheism",
+        "stance_feminist",
+        "stance_hillary",
+        "match",
+        "goemo_soft",
+        "reason",
+        "formality",
+        "politeness",
+        "strategyqa",
+        "vitaminc",
+        "ruletaker",
+        "proofwriter",
+        "folio",
+        "logiqa",
+        "tracie",
+        "temporal_nli",
+        "piqa",
+        "siqa",
+        "clutrr",
+        "gsm8k",
+        "svamp",
+        "aqua"
+      ],
+      "holdout": [
+        "probe",
+        "bbh",
+        "cola",
+        "wic",
+        "subj",
+        "spam",
+        "counterfactual",
+        "cb",
+        "arc_challenge",
+        "stance_climate",
+        "trec",
+        "sst5",
+        "fin_sentiment",
+        "arc_easy",
+        "newsgroups"
+      ],
+      "steps": 12118,
+      "train_examples": 158246,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 16,
+        "loss": "mix",
+        "max_per_task": 3000,
+        "epochs": 1,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {
+        "ag_news": {
+          "n": 300,
+          "acc": 0.9233333333333333,
+          "nll": 0.23956205062784391,
+          "brier": 0.12588630912802798,
+          "ece": 0.03565445333719247,
+          "mean_conf": 0.9105748584866524,
+          "cov@0.5": 0.9833333333333333,
+          "acc@0.5": 0.9288135593220339,
+          "cov@0.7": 0.88,
+          "acc@0.7": 0.9545454545454546,
+          "cov@0.9": 0.77,
+          "acc@0.9": 0.9653679653679653
+        },
+        "boolq": {
+          "n": 300,
+          "acc": 0.8766666666666667,
+          "nll": 0.33278360864483053,
+          "brier": 0.19905847383131203,
+          "ece": 0.048838321963946066,
+          "mean_conf": 0.8610335459311803,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.8766666666666667,
+          "cov@0.7": 0.8866666666666667,
+          "acc@0.7": 0.9060150375939849,
+          "cov@0.9": 0.4666666666666667,
+          "acc@0.9": 0.9642857142857143
+        },
+        "commonsense_qa": {
+          "n": 300,
+          "acc": 0.7866666666666666,
+          "nll": 0.5812667453454432,
+          "brier": 0.2960317921716136,
+          "ece": 0.05657371540864312,
+          "mean_conf": 0.7944276158014933,
+          "cov@0.5": 0.8933333333333333,
+          "acc@0.5": 0.832089552238806,
+          "cov@0.7": 0.7,
+          "acc@0.7": 0.9047619047619048,
+          "cov@0.9": 0.42333333333333334,
+          "acc@0.9": 0.952755905511811
+        },
+        "copa": {
+          "n": 200,
+          "acc": 0.93,
+          "nll": 0.17489093500636219,
+          "brier": 0.10265215895968784,
+          "ece": 0.03722041517496112,
+          "mean_conf": 0.8977384361624717,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.93,
+          "cov@0.7": 0.9,
+          "acc@0.7": 0.9611111111111111,
+          "cov@0.9": 0.69,
+          "acc@0.9": 1.0
+        },
+        "dbpedia": {
+          "n": 300,
+          "acc": 0.97,
+          "nll": 0.12029726431718998,
+          "brier": 0.0467073057529959,
+          "ece": 0.011912721196810366,
+          "mean_conf": 0.9797261367241542,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.97,
+          "cov@0.7": 0.9866666666666667,
+          "acc@0.7": 0.9797297297297297,
+          "cov@0.9": 0.9766666666666667,
+          "acc@0.9": 0.9829351535836177
+        },
+        "emotion": {
+          "n": 300,
+          "acc": 0.7766666666666666,
+          "nll": 0.6417595775959655,
+          "brier": 0.31279347659514983,
+          "ece": 0.07161758591731388,
+          "mean_conf": 0.8046423467993736,
+          "cov@0.5": 0.92,
+          "acc@0.5": 0.8043478260869565,
+          "cov@0.7": 0.7233333333333334,
+          "acc@0.7": 0.9032258064516129,
+          "cov@0.9": 0.4266666666666667,
+          "acc@0.9": 0.9609375
+        },
+        "facts": {
+          "n": 54,
+          "acc": 0.9629629629629629,
+          "nll": 0.10233185567308967,
+          "brier": 0.0580309563229181,
+          "ece": 0.03901768503365692,
+          "mean_conf": 0.985795874286581,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.9629629629629629,
+          "cov@0.7": 1.0,
+          "acc@0.7": 0.9629629629629629,
+          "cov@0.9": 0.9444444444444444,
+          "acc@0.9": 1.0
+        },
+        "fits": {
+          "n": 214,
+          "acc": 0.8598130841121495,
+          "nll": 0.4789074549575723,
+          "brier": 0.09971697154134483,
+          "ece": 0.08341867177285883,
+          "mean_conf": 0.818978947456752,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.8598130841121495,
+          "cov@0.7": 0.8037383177570093,
+          "acc@0.7": 0.936046511627907,
+          "cov@0.9": 0.3317757009345794,
+          "acc@0.9": 1.0
+        },
+        "imdb": {
+          "n": 300,
+          "acc": 0.95,
+          "nll": 0.1449115194526972,
+          "brier": 0.07315853850730186,
+          "ece": 0.028431474765141816,
+          "mean_conf": 0.9752282838026682,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.95,
+          "cov@0.7": 0.9833333333333333,
+          "acc@0.7": 0.9627118644067797,
+          "cov@0.9": 0.95,
+          "acc@0.9": 0.9754385964912281
+        },
+        "jailbreak": {
+          "n": 200,
+          "acc": 0.98,
+          "nll": 0.07703215111456595,
+          "brier": 0.03552832502947578,
+          "ece": 0.014975480735301977,
+          "mean_conf": 0.972049820125103,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.98,
+          "cov@0.7": 0.995,
+          "acc@0.7": 0.9798994974874372,
+          "cov@0.9": 0.965,
+          "acc@0.9": 0.9896373056994818
+        },
+        "mnli": {
+          "n": 300,
+          "acc": 0.8466666666666667,
+          "nll": 0.44322894987837685,
+          "brier": 0.24040505319501904,
+          "ece": 0.05213867117961251,
+          "mean_conf": 0.8880125307043394,
+          "cov@0.5": 0.9933333333333333,
+          "acc@0.5": 0.8489932885906041,
+          "cov@0.7": 0.9,
+          "acc@0.7": 0.8814814814814815,
+          "cov@0.9": 0.64,
+          "acc@0.9": 0.9427083333333334
+        },
+        "mrpc": {
+          "n": 200,
+          "acc": 0.83,
+          "nll": 0.3825738710855869,
+          "brier": 0.2414001594036636,
+          "ece": 0.0455596360564232,
+          "mean_conf": 0.8177212104201317,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.83,
+          "cov@0.7": 0.745,
+          "acc@0.7": 0.9060402684563759,
+          "cov@0.9": 0.415,
+          "acc@0.9": 0.9518072289156626
+        },
+        "openbookqa": {
+          "n": 300,
+          "acc": 0.8166666666666667,
+          "nll": 0.4884890429825752,
+          "brier": 0.26006302091594946,
+          "ece": 0.06033645391464233,
+          "mean_conf": 0.8382209448019663,
+          "cov@0.5": 0.9266666666666666,
+          "acc@0.5": 0.8525179856115108,
+          "cov@0.7": 0.7966666666666666,
+          "acc@0.7": 0.899581589958159,
+          "cov@0.9": 0.52,
+          "acc@0.9": 0.9743589743589743
+        },
+        "paws": {
+          "n": 300,
+          "acc": 0.93,
+          "nll": 0.2171823905014709,
+          "brier": 0.11854569746905995,
+          "ece": 0.028286268909772223,
+          "mean_conf": 0.9089529289801915,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.93,
+          "cov@0.7": 0.9166666666666666,
+          "acc@0.7": 0.9490909090909091,
+          "cov@0.9": 0.7366666666666667,
+          "acc@0.9": 0.9638009049773756
+        },
+        "qnli": {
+          "n": 300,
+          "acc": 0.88,
+          "nll": 0.29461164414824,
+          "brier": 0.1746824483593742,
+          "ece": 0.03399742662906641,
+          "mean_conf": 0.9018590325117111,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.88,
+          "cov@0.7": 0.92,
+          "acc@0.7": 0.9166666666666666,
+          "cov@0.9": 0.6933333333333334,
+          "acc@0.9": 0.9423076923076923
+        },
+        "read": {
+          "n": 300,
+          "acc": 1.0,
+          "nll": 0.013479419595217349,
+          "brier": 0.0006464961621803141,
+          "ece": 0.01327262739340459,
+          "mean_conf": 0.9867273726065954,
+          "cov@0.5": 1.0,
+          "acc@0.5": 1.0,
+          "cov@0.7": 1.0,
+          "acc@0.7": 1.0,
+          "cov@0.9": 0.9966666666666667,
+          "acc@0.9": 1.0
+        },
+        "rte": {
+          "n": 200,
+          "acc": 0.875,
+          "nll": 0.2619901953560894,
+          "brier": 0.1655817686780098,
+          "ece": 0.07393659174442292,
+          "mean_conf": 0.8983567506074905,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.875,
+          "cov@0.7": 0.915,
+          "acc@0.7": 0.912568306010929,
+          "cov@0.9": 0.69,
+          "acc@0.9": 0.9855072463768116
+        },
+        "sciq": {
+          "n": 300,
+          "acc": 0.9833333333333333,
+          "nll": 0.0542967460035165,
+          "brier": 0.02836314450778487,
+          "ece": 0.013889081676801088,
+          "mean_conf": 0.9804451249043147,
+          "cov@0.5": 0.9966666666666667,
+          "acc@0.5": 0.9866220735785953,
+          "cov@0.7": 0.99,
+          "acc@0.7": 0.9865319865319865,
+          "cov@0.9": 0.95,
+          "acc@0.9": 0.9929824561403509
+        },
+        "sst2": {
+          "n": 300,
+          "acc": 0.94,
+          "nll": 0.15217635479920968,
+          "brier": 0.08669829778686115,
+          "ece": 0.02306454201539363,
+          "mean_conf": 0.9453726333379745,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.94,
+          "cov@0.7": 0.9633333333333334,
+          "acc@0.7": 0.9584775086505191,
+          "cov@0.9": 0.82,
+          "acc@0.9": 0.991869918699187
+        },
+        "swag": {
+          "n": 300,
+          "acc": 0.7666666666666667,
+          "nll": 0.7234309196216375,
+          "brier": 0.3695786167661231,
+          "ece": 0.06823263516028721,
+          "mean_conf": 0.7794082881013552,
+          "cov@0.5": 0.91,
+          "acc@0.5": 0.7875457875457875,
+          "cov@0.7": 0.69,
+          "acc@0.7": 0.8405797101449275,
+          "cov@0.9": 0.3333333333333333,
+          "acc@0.9": 0.9
+        },
+        "tweet_emoji": {
+          "n": 300,
+          "acc": 0.24333333333333335,
+          "nll": 2.5063694445877984,
+          "brier": 0.83829218239292,
+          "ece": 0.08510932529966037,
+          "mean_conf": 0.2803319871922334,
+          "cov@0.5": 0.12,
+          "acc@0.5": 0.8055555555555556,
+          "cov@0.7": 0.07666666666666666,
+          "acc@0.7": 0.9130434782608695,
+          "cov@0.9": 0.0033333333333333335,
+          "acc@0.9": 1.0
+        },
+        "tweet_hate": {
+          "n": 300,
+          "acc": 0.7166666666666667,
+          "nll": 0.5309249597827629,
+          "brier": 0.36006795343339115,
+          "ece": 0.09840704739093784,
+          "mean_conf": 0.8124431739250819,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.7166666666666667,
+          "cov@0.7": 0.7866666666666666,
+          "acc@0.7": 0.7966101694915254,
+          "cov@0.9": 0.32666666666666666,
+          "acc@0.9": 0.9285714285714286
+        },
+        "tweet_irony": {
+          "n": 300,
+          "acc": 0.71,
+          "nll": 0.5534378644031082,
+          "brier": 0.37644439644653505,
+          "ece": 0.03676844378312428,
+          "mean_conf": 0.7351771769920985,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.71,
+          "cov@0.7": 0.5866666666666667,
+          "acc@0.7": 0.8068181818181818,
+          "cov@0.9": 0.12333333333333334,
+          "acc@0.9": 0.972972972972973
+        },
+        "tweet_offensive": {
+          "n": 300,
+          "acc": 0.7866666666666666,
+          "nll": 0.47620018437820133,
+          "brier": 0.3128923872633037,
+          "ece": 0.05815195361773175,
+          "mean_conf": 0.8035398570696513,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.7866666666666666,
+          "cov@0.7": 0.76,
+          "acc@0.7": 0.8421052631578947,
+          "cov@0.9": 0.3333333333333333,
+          "acc@0.9": 0.93
+        },
+        "tweet_sentiment": {
+          "n": 300,
+          "acc": 0.7266666666666667,
+          "nll": 0.6050564579947201,
+          "brier": 0.36215817406899686,
+          "ece": 0.07843278278907143,
+          "mean_conf": 0.7284325797359149,
+          "score_mae": 0.3644068883561219,
+          "cov@0.5": 0.96,
+          "acc@0.5": 0.7291666666666666,
+          "cov@0.7": 0.5333333333333333,
+          "acc@0.7": 0.875,
+          "cov@0.9": 0.16666666666666666,
+          "acc@0.9": 0.92
+        },
+        "yahoo": {
+          "n": 300,
+          "acc": 0.7166666666666667,
+          "nll": 0.8738976731967112,
+          "brier": 0.3950192633049044,
+          "ece": 0.092816769828399,
+          "mean_conf": 0.8022240548829238,
+          "cov@0.5": 0.9133333333333333,
+          "acc@0.5": 0.7627737226277372,
+          "cov@0.7": 0.7466666666666667,
+          "acc@0.7": 0.8348214285714286,
+          "cov@0.9": 0.4166666666666667,
+          "acc@0.9": 0.92
+        },
+        "yelp": {
+          "n": 300,
+          "acc": 0.71,
+          "nll": 0.731893961859198,
+          "brier": 0.42142868253511284,
+          "ece": 0.09561944127082822,
+          "mean_conf": 0.7473884936173757,
+          "score_mae": 0.3743417001541093,
+          "cov@0.5": 0.9533333333333334,
+          "acc@0.5": 0.7237762237762237,
+          "cov@0.7": 0.6266666666666667,
+          "acc@0.7": 0.7712765957446809,
+          "cov@0.9": 0.19,
+          "acc@0.9": 0.9649122807017544
+        },
+        "anli": {
+          "n": 300,
+          "acc": 0.5366666666666666,
+          "nll": 1.1646257930212216,
+          "brier": 0.675594648318733,
+          "ece": 0.223523634771506,
+          "mean_conf": 0.7592115387320518,
+          "cov@0.5": 0.9433333333333334,
+          "acc@0.5": 0.5547703180212014,
+          "cov@0.7": 0.65,
+          "acc@0.7": 0.558974358974359,
+          "cov@0.9": 0.23,
+          "acc@0.9": 0.5507246376811594
+        },
+        "winogrande": {
+          "n": 300,
+          "acc": 0.7966666666666666,
+          "nll": 0.46445836318766226,
+          "brier": 0.3013142673360519,
+          "ece": 0.07452347179253899,
+          "mean_conf": 0.8619708905617396,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.7966666666666666,
+          "cov@0.7": 0.87,
+          "acc@0.7": 0.8275862068965517,
+          "cov@0.9": 0.55,
+          "acc@0.9": 0.9030303030303031
+        },
+        "hellaswag": {
+          "n": 300,
+          "acc": 0.8566666666666667,
+          "nll": 0.3762581950318828,
+          "brier": 0.19513976820616458,
+          "ece": 0.05684188375870386,
+          "mean_conf": 0.8465098922451337,
+          "cov@0.5": 0.9366666666666666,
+          "acc@0.5": 0.9074733096085409,
+          "cov@0.7": 0.7833333333333333,
+          "acc@0.7": 0.9446808510638298,
+          "cov@0.9": 0.5533333333333333,
+          "acc@0.9": 0.9819277108433735
+        },
+        "race": {
+          "n": 300,
+          "acc": 0.7933333333333333,
+          "nll": 0.5311142158904583,
+          "brier": 0.2739426106933522,
+          "ece": 0.0815848172704379,
+          "mean_conf": 0.8720341417193412,
+          "cov@0.5": 0.93,
+          "acc@0.5": 0.8422939068100358,
+          "cov@0.7": 0.8466666666666667,
+          "acc@0.7": 0.8818897637795275,
+          "cov@0.9": 0.6566666666666666,
+          "acc@0.9": 0.934010152284264
+        },
+        "scitail": {
+          "n": 300,
+          "acc": 0.97,
+          "nll": 0.1026227491797548,
+          "brier": 0.05202258655692825,
+          "ece": 0.028236998518308055,
+          "mean_conf": 0.950906420747439,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.97,
+          "cov@0.7": 0.9833333333333333,
+          "acc@0.7": 0.9728813559322034,
+          "cov@0.9": 0.8466666666666667,
+          "acc@0.9": 0.9921259842519685
+        },
+        "qqp": {
+          "n": 300,
+          "acc": 0.88,
+          "nll": 0.2769447782152307,
+          "brier": 0.17137568112639565,
+          "ece": 0.045283984939257296,
+          "mean_conf": 0.8734472642342249,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.88,
+          "cov@0.7": 0.8533333333333334,
+          "acc@0.7": 0.93359375,
+          "cov@0.9": 0.5766666666666667,
+          "acc@0.9": 0.9710982658959537
+        },
+        "stsb": {
+          "n": 287,
+          "acc": 0.6167247386759582,
+          "nll": 1.0364299467274243,
+          "brier": 0.2999134000942182,
+          "ece": 0.07439108956150894,
+          "mean_conf": 0.5469927265461314,
+          "score_mae": 0.5417481492766754,
+          "cov@0.5": 0.6306620209059234,
+          "acc@0.5": 0.6850828729281768,
+          "cov@0.7": 0.09407665505226481,
+          "acc@0.7": 0.8518518518518519,
+          "cov@0.9": 0.0,
+          "acc@0.9": NaN
+        },
+        "toxic": {
+          "n": 300,
+          "acc": 0.84,
+          "nll": 0.34017511751074586,
+          "brier": 0.21106674918538404,
+          "ece": 0.04502722958723705,
+          "mean_conf": 0.8586645072698593,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.84,
+          "cov@0.7": 0.88,
+          "acc@0.7": 0.8977272727272727,
+          "cov@0.9": 0.52,
+          "acc@0.9": 0.9615384615384616
+        },
+        "stance_abortion": {
+          "n": 66,
+          "acc": 0.8636363636363636,
+          "nll": 0.44344493585724276,
+          "brier": 0.23815874340787538,
+          "ece": 0.16279316354881634,
+          "mean_conf": 0.7459786872972142,
+          "cov@0.5": 0.8636363636363636,
+          "acc@0.5": 0.8947368421052632,
+          "cov@0.7": 0.5909090909090909,
+          "acc@0.7": 0.9230769230769231,
+          "cov@0.9": 0.22727272727272727,
+          "acc@0.9": 1.0
+        },
+        "stance_atheism": {
+          "n": 52,
+          "acc": 0.7115384615384616,
+          "nll": 0.6112422593374254,
+          "brier": 0.37005177211571516,
+          "ece": 0.12039663585332726,
+          "mean_conf": 0.7864993226069671,
+          "cov@0.5": 0.9423076923076923,
+          "acc@0.5": 0.7346938775510204,
+          "cov@0.7": 0.6538461538461539,
+          "acc@0.7": 0.8235294117647058,
+          "cov@0.9": 0.3269230769230769,
+          "acc@0.9": 1.0
+        },
+        "stance_feminist": {
+          "n": 67,
+          "acc": 0.7313432835820896,
+          "nll": 0.7303124558726009,
+          "brier": 0.4342147904610734,
+          "ece": 0.1691147155726134,
+          "mean_conf": 0.7381730840277316,
+          "cov@0.5": 0.9402985074626866,
+          "acc@0.5": 0.746031746031746,
+          "cov@0.7": 0.5522388059701493,
+          "acc@0.7": 0.7027027027027027,
+          "cov@0.9": 0.19402985074626866,
+          "acc@0.9": 0.8461538461538461
+        },
+        "stance_hillary": {
+          "n": 69,
+          "acc": 0.7246376811594203,
+          "nll": 0.6620274924672167,
+          "brier": 0.3935280964850548,
+          "ece": 0.09731849552928537,
+          "mean_conf": 0.7686952758526456,
+          "cov@0.5": 0.9565217391304348,
+          "acc@0.5": 0.7424242424242424,
+          "cov@0.7": 0.6666666666666666,
+          "acc@0.7": 0.8043478260869565,
+          "cov@0.9": 0.18840579710144928,
+          "acc@0.9": 0.7692307692307693
+        },
+        "match": {
+          "n": 300,
+          "acc": 0.99,
+          "nll": 0.03724694484885282,
+          "brier": 0.016051036461267973,
+          "ece": 0.009268070856730138,
+          "mean_conf": 0.9877331558863321,
+          "cov@0.5": 0.9966666666666667,
+          "acc@0.5": 0.9933110367892977,
+          "cov@0.7": 0.9966666666666667,
+          "acc@0.7": 0.9933110367892977,
+          "cov@0.9": 0.99,
+          "acc@0.9": 0.9932659932659933
+        },
+        "reason": {
+          "n": 300,
+          "acc": 0.9466666666666667,
+          "nll": 0.17577242359452994,
+          "brier": 0.09125921721706085,
+          "ece": 0.04738917231559757,
+          "mean_conf": 0.9024230941136678,
+          "cov@0.5": 0.9833333333333333,
+          "acc@0.5": 0.9559322033898305,
+          "cov@0.7": 0.88,
+          "acc@0.7": 0.9772727272727273,
+          "cov@0.9": 0.7733333333333333,
+          "acc@0.9": 0.9913793103448276
+        },
+        "formality": {
+          "n": 300,
+          "acc": 0.6066666666666667,
+          "nll": 0.9962800608084179,
+          "brier": 0.25308033830347887,
+          "ece": 0.07329747378826143,
+          "mean_conf": 0.5359433833758036,
+          "score_mae": 0.47120167226336584,
+          "cov@0.5": 0.6866666666666666,
+          "acc@0.5": 0.6699029126213593,
+          "cov@0.7": 0.0033333333333333335,
+          "acc@0.7": 1.0,
+          "cov@0.9": 0.0,
+          "acc@0.9": NaN
+        },
+        "politeness": {
+          "n": 300,
+          "acc": 0.8633333333333333,
+          "nll": 0.3753622173094621,
+          "brier": 0.18929635618156296,
+          "ece": 0.026931450863679252,
+          "mean_conf": 0.8615820496280988,
+          "score_mae": 0.21370556724568207,
+          "cov@0.5": 0.9633333333333334,
+          "acc@0.5": 0.8858131487889274,
+          "cov@0.7": 0.8266666666666667,
+          "acc@0.7": 0.9435483870967742,
+          "cov@0.9": 0.5766666666666667,
+          "acc@0.9": 0.9826589595375722
+        },
+        "strategyqa": {
+          "n": 200,
+          "acc": 0.67,
+          "nll": 0.5862248309774405,
+          "brier": 0.40554963519042464,
+          "ece": 0.05940887540578843,
+          "mean_conf": 0.6801840284466744,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.67,
+          "cov@0.7": 0.385,
+          "acc@0.7": 0.8051948051948052,
+          "cov@0.9": 0.07,
+          "acc@0.9": 1.0
+        },
+        "vitaminc": {
+          "n": 300,
+          "acc": 0.82,
+          "nll": 0.5214132864032636,
+          "brier": 0.2848927857191196,
+          "ece": 0.03640172024567924,
+          "mean_conf": 0.8344497634967168,
+          "cov@0.5": 0.9666666666666667,
+          "acc@0.5": 0.8275862068965517,
+          "cov@0.7": 0.83,
+          "acc@0.7": 0.8634538152610441,
+          "cov@0.9": 0.4533333333333333,
+          "acc@0.9": 0.9338235294117647
+        },
+        "ruletaker": {
+          "n": 300,
+          "acc": 0.8066666666666666,
+          "nll": 0.3848169850830573,
+          "brier": 0.25008683944182986,
+          "ece": 0.030118134220441215,
+          "mean_conf": 0.812346151471138,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.8066666666666666,
+          "cov@0.7": 0.7133333333333334,
+          "acc@0.7": 0.9018691588785047,
+          "cov@0.9": 0.43666666666666665,
+          "acc@0.9": 0.9618320610687023
+        },
+        "proofwriter": {
+          "n": 300,
+          "acc": 0.7933333333333333,
+          "nll": 0.4490651794863921,
+          "brier": 0.2722325912094693,
+          "ece": 0.06963174422581991,
+          "mean_conf": 0.8091626433531444,
+          "cov@0.5": 0.98,
+          "acc@0.5": 0.7993197278911565,
+          "cov@0.7": 0.7633333333333333,
+          "acc@0.7": 0.8777292576419214,
+          "cov@0.9": 0.37333333333333335,
+          "acc@0.9": 0.9910714285714286
+        },
+        "folio": {
+          "n": 200,
+          "acc": 0.625,
+          "nll": 0.7944451580787752,
+          "brier": 0.4657608423671937,
+          "ece": 0.11621890529990193,
+          "mean_conf": 0.6586561058461666,
+          "cov@0.5": 0.835,
+          "acc@0.5": 0.688622754491018,
+          "cov@0.7": 0.375,
+          "acc@0.7": 0.8533333333333334,
+          "cov@0.9": 0.06,
+          "acc@0.9": 0.75
+        },
+        "logiqa": {
+          "n": 300,
+          "acc": 0.5833333333333334,
+          "nll": 0.6570202462503513,
+          "brier": 0.46499107764281516,
+          "ece": 0.061852243741353355,
+          "mean_conf": 0.6316624116897583,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.5833333333333334,
+          "cov@0.7": 0.23666666666666666,
+          "acc@0.7": 0.7323943661971831,
+          "cov@0.9": 0.013333333333333334,
+          "acc@0.9": 0.75
+        },
+        "tracie": {
+          "n": 200,
+          "acc": 0.77,
+          "nll": 0.5158300579084877,
+          "brier": 0.3394760687221474,
+          "ece": 0.07212192535400391,
+          "mean_conf": 0.70532983481884,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.77,
+          "cov@0.7": 0.49,
+          "acc@0.7": 0.8673469387755102,
+          "cov@0.9": 0.03,
+          "acc@0.9": 1.0
+        },
+        "temporal_nli": {
+          "n": 300,
+          "acc": 0.8633333333333333,
+          "nll": 0.3496475835064699,
+          "brier": 0.20055396242126278,
+          "ece": 0.07610381027062735,
+          "mean_conf": 0.7972539271910986,
+          "cov@0.5": 0.9866666666666667,
+          "acc@0.5": 0.8614864864864865,
+          "cov@0.7": 0.7666666666666667,
+          "acc@0.7": 0.9217391304347826,
+          "cov@0.9": 0.26666666666666666,
+          "acc@0.9": 1.0
+        },
+        "piqa": {
+          "n": 300,
+          "acc": 0.8166666666666667,
+          "nll": 0.36417624160320555,
+          "brier": 0.23331890879921213,
+          "ece": 0.038980930248896296,
+          "mean_conf": 0.8269510519504547,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.8166666666666667,
+          "cov@0.7": 0.7633333333333333,
+          "acc@0.7": 0.9126637554585153,
+          "cov@0.9": 0.43666666666666665,
+          "acc@0.9": 0.9618320610687023
+        },
+        "siqa": {
+          "n": 300,
+          "acc": 0.8233333333333334,
+          "nll": 0.4337533873205736,
+          "brier": 0.24543190225066103,
+          "ece": 0.034700063069661474,
+          "mean_conf": 0.8447816316286723,
+          "cov@0.5": 0.9633333333333334,
+          "acc@0.5": 0.8408304498269896,
+          "cov@0.7": 0.7833333333333333,
+          "acc@0.7": 0.9148936170212766,
+          "cov@0.9": 0.5533333333333333,
+          "acc@0.9": 0.9457831325301205
+        },
+        "clutrr": {
+          "n": 300,
+          "acc": 0.6133333333333333,
+          "nll": 0.9127365656335344,
+          "brier": 0.4770832175168861,
+          "ece": 0.04643365234136578,
+          "mean_conf": 0.5962887792785962,
+          "cov@0.5": 0.72,
+          "acc@0.5": 0.6944444444444444,
+          "cov@0.7": 0.21333333333333335,
+          "acc@0.7": 0.9375,
+          "cov@0.9": 0.08333333333333333,
+          "acc@0.9": 1.0
+        },
+        "gsm8k": {
+          "n": 300,
+          "acc": 0.74,
+          "nll": 0.589682709113467,
+          "brier": 0.32688956856451207,
+          "ece": 0.06247660587231318,
+          "mean_conf": 0.7070875727136929,
+          "cov@0.5": 0.7866666666666666,
+          "acc@0.5": 0.826271186440678,
+          "cov@0.7": 0.53,
+          "acc@0.7": 0.9308176100628931,
+          "cov@0.9": 0.25,
+          "acc@0.9": 1.0
+        },
+        "svamp": {
+          "n": 100,
+          "acc": 0.63,
+          "nll": 0.8085613738920417,
+          "brier": 0.4651299440112998,
+          "ece": 0.12156755775213245,
+          "mean_conf": 0.601172327697277,
+          "cov@0.5": 0.68,
+          "acc@0.5": 0.75,
+          "cov@0.7": 0.32,
+          "acc@0.7": 0.875,
+          "cov@0.9": 0.11,
+          "acc@0.9": 1.0
+        },
+        "aqua": {
+          "n": 253,
+          "acc": 0.383399209486166,
+          "nll": 1.467006593256583,
+          "brier": 0.7354768193410129,
+          "ece": 0.05388568453637978,
+          "mean_conf": 0.3621667535173092,
+          "cov@0.5": 0.09486166007905138,
+          "acc@0.5": 0.6666666666666666,
+          "cov@0.7": 0.019762845849802372,
+          "acc@0.7": 0.4,
+          "cov@0.9": 0.0,
+          "acc@0.9": NaN
+        }
+      },
+      "unseen_test_uncalibrated": {
+        "probe": {
+          "n": 97,
+          "acc": 0.8969072164948454,
+          "nll": 0.20691493444770823,
+          "brier": 0.12394450383288905,
+          "ece": 0.06851052070401381,
+          "mean_conf": 0.9103295477395205,
+          "score_mae": 0.18648642087646294,
+          "cov@0.5": 0.9896907216494846,
+          "acc@0.5": 0.90625,
+          "cov@0.7": 0.9072164948453608,
+          "acc@0.7": 0.9545454545454546,
+          "cov@0.9": 0.7628865979381443,
+          "acc@0.9": 0.9864864864864865,
+          "families": {
+            "desc": [
+              13,
+              15
+            ],
+            "negation": [
+              10,
+              10
+            ],
+            "logic": [
+              14,
+              17
+            ],
+            "score": [
+              10,
+              12
+            ],
+            "json": [
+              6,
+              6
+            ],
+            "taxonomy": [
+              14,
+              14
+            ],
+            "plausible": [
+              4,
+              4
+            ],
+            "twist": [
+              3,
+              3
+            ],
+            "time": [
+              4,
+              5
+            ],
+            "intent": [
+              2,
+              3
+            ],
+            "compare": [
+              3,
+              4
+            ],
+            "criteria": [
+              4,
+              4
+            ]
+          }
+        },
+        "bbh": {
+          "n": 1000,
+          "acc": 0.517,
+          "nll": 1.108523054891869,
+          "brier": 0.6035352775259032,
+          "ece": 0.07363909149914981,
+          "mean_conf": 0.5724366051629186,
+          "cov@0.5": 0.64,
+          "acc@0.5": 0.6140625,
+          "cov@0.7": 0.273,
+          "acc@0.7": 0.6959706959706959,
+          "cov@0.9": 0.071,
+          "acc@0.9": 0.7887323943661971
+        },
+        "cola": {
+          "n": 1000,
+          "acc": 0.75,
+          "nll": 0.5104252819118784,
+          "brier": 0.33587838373719203,
+          "ece": 0.0569284417629242,
+          "mean_conf": 0.7190280594825744,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.75,
+          "cov@0.7": 0.58,
+          "acc@0.7": 0.8448275862068966,
+          "cov@0.9": 0.013,
+          "acc@0.9": 1.0
+        },
+        "wic": {
+          "n": 638,
+          "acc": 0.5909090909090909,
+          "nll": 0.69998100116263,
+          "brier": 0.4982001306496651,
+          "ece": 0.09791882723850147,
+          "mean_conf": 0.6873193471969855,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.5909090909090909,
+          "cov@0.7": 0.44357366771159873,
+          "acc@0.7": 0.6537102473498233,
+          "cov@0.9": 0.0219435736677116,
+          "acc@0.9": 0.7142857142857143
+        },
+        "subj": {
+          "n": 1000,
+          "acc": 0.682,
+          "nll": 0.5917236264696343,
+          "brier": 0.4104362382942676,
+          "ece": 0.1212564522027969,
+          "mean_conf": 0.7977630772590637,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.682,
+          "cov@0.7": 0.733,
+          "acc@0.7": 0.7517053206002728,
+          "cov@0.9": 0.306,
+          "acc@0.9": 0.9215686274509803
+        },
+        "spam": {
+          "n": 1000,
+          "acc": 0.749,
+          "nll": 0.47590900971852373,
+          "brier": 0.32175875624611966,
+          "ece": 0.10708969771862033,
+          "mean_conf": 0.8200223511457443,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.749,
+          "cov@0.7": 0.816,
+          "acc@0.7": 0.803921568627451,
+          "cov@0.9": 0.353,
+          "acc@0.9": 0.9773371104815864
+        },
+        "counterfactual": {
+          "n": 1000,
+          "acc": 0.82,
+          "nll": 0.4564823726332652,
+          "brier": 0.2849392428188435,
+          "ece": 0.10961719477176668,
+          "mean_conf": 0.7216410273313523,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.82,
+          "cov@0.7": 0.617,
+          "acc@0.7": 0.9141004862236629,
+          "cov@0.9": 0.01,
+          "acc@0.9": 0.8
+        },
+        "cb": {
+          "n": 56,
+          "acc": 0.875,
+          "nll": 0.3821062575477204,
+          "brier": 0.19684515831431662,
+          "ece": 0.11659146206719534,
+          "mean_conf": 0.8077813791377204,
+          "cov@0.5": 0.9821428571428571,
+          "acc@0.5": 0.8909090909090909,
+          "cov@0.7": 0.75,
+          "acc@0.7": 0.9761904761904762,
+          "cov@0.9": 0.35714285714285715,
+          "acc@0.9": 1.0
+        },
+        "arc_challenge": {
+          "n": 1000,
+          "acc": 0.79,
+          "nll": 0.5784666197380147,
+          "brier": 0.3084519084178791,
+          "ece": 0.04053584739565846,
+          "mean_conf": 0.8078768512308597,
+          "cov@0.5": 0.897,
+          "acc@0.5": 0.8249721293199554,
+          "cov@0.7": 0.719,
+          "acc@0.7": 0.8873435326842837,
+          "cov@0.9": 0.46,
+          "acc@0.9": 0.9586956521739131
+        },
+        "stance_climate": {
+          "n": 169,
+          "acc": 0.7100591715976331,
+          "nll": 0.8004923563430824,
+          "brier": 0.43626973420217074,
+          "ece": 0.08912066789068414,
+          "mean_conf": 0.7112431993498605,
+          "cov@0.5": 0.863905325443787,
+          "acc@0.5": 0.726027397260274,
+          "cov@0.7": 0.5502958579881657,
+          "acc@0.7": 0.8709677419354839,
+          "cov@0.9": 0.1301775147928994,
+          "acc@0.9": 0.9545454545454546
+        },
+        "trec": {
+          "n": 500,
+          "acc": 0.728,
+          "nll": 0.7789360883597788,
+          "brier": 0.3915759426435931,
+          "ece": 0.055817239046096784,
+          "mean_conf": 0.7733147183656692,
+          "cov@0.5": 0.89,
+          "acc@0.5": 0.7797752808988764,
+          "cov@0.7": 0.668,
+          "acc@0.7": 0.8323353293413174,
+          "cov@0.9": 0.362,
+          "acc@0.9": 0.8895027624309392
+        },
+        "sst5": {
+          "n": 1000,
+          "acc": 0.569,
+          "nll": 0.9569207478808627,
+          "brier": 0.5500160496510278,
+          "ece": 0.051933742076158515,
+          "mean_conf": 0.5590169258415699,
+          "score_mae": 0.5127075865020743,
+          "cov@0.5": 0.684,
+          "acc@0.5": 0.6198830409356725,
+          "cov@0.7": 0.117,
+          "acc@0.7": 0.7094017094017094,
+          "cov@0.9": 0.005,
+          "acc@0.9": 0.6
+        },
+        "fin_sentiment": {
+          "n": 1000,
+          "acc": 0.821,
+          "nll": 0.45971791800530354,
+          "brier": 0.26788516663028156,
+          "ece": 0.06331040862202642,
+          "mean_conf": 0.7591086620986461,
+          "cov@0.5": 0.981,
+          "acc@0.5": 0.8287461773700305,
+          "cov@0.7": 0.684,
+          "acc@0.7": 0.9078947368421053,
+          "cov@0.9": 0.139,
+          "acc@0.9": 0.9640287769784173
+        },
+        "arc_easy": {
+          "n": 1000,
+          "acc": 0.893,
+          "nll": 0.30681700111668636,
+          "brier": 0.16114664739945947,
+          "ece": 0.017386740416288377,
+          "mean_conf": 0.8903764767348766,
+          "cov@0.5": 0.959,
+          "acc@0.5": 0.9113660062565172,
+          "cov@0.7": 0.864,
+          "acc@0.7": 0.9444444444444444,
+          "cov@0.9": 0.704,
+          "acc@0.9": 0.9758522727272727
+        },
+        "newsgroups": {
+          "n": 1000,
+          "acc": 0.645,
+          "nll": 1.2438821946531702,
+          "brier": 0.4654980387667391,
+          "ece": 0.0510300065651536,
+          "mean_conf": 0.6572305353507399,
+          "cov@0.5": 0.683,
+          "acc@0.5": 0.8272327964860908,
+          "cov@0.7": 0.5,
+          "acc@0.7": 0.9,
+          "cov@0.9": 0.246,
+          "acc@0.9": 0.9878048780487805
+        }
+      },
+      "summary": {
+        "mean_acc": 0.7922933763477235,
+        "mean_ece": 0.06318428710662419
+      },
+      "history": [],
+      "temperature": 1.0194810628890991,
+      "temperature_by_k": {
+        "2-2": 0.9909,
+        "3-5": 1.0355,
+        "6-20": 1.0193
+      },
+      "calibration": {
+        "before": {
+          "n": 10608,
+          "acc": 0.7931749622926093,
+          "nll": 0.5081595549816602,
+          "brier": 0.26480504720767034,
+          "ece": 0.010474447911776612,
+          "mean_conf": 0.8014159178505982,
+          "score_mae": 0.3837145484503708,
+          "cov@0.5": 0.9200603318250377,
+          "acc@0.5": 0.8326844262295082,
+          "cov@0.7": 0.7074849170437406,
+          "acc@0.7": 0.9077948034643571,
+          "cov@0.9": 0.4572963800904977,
+          "acc@0.9": 0.9647495361781077
+        },
+        "after": {
+          "n": 10608,
+          "acc": 0.7931749622926093,
+          "nll": 0.5079944898473756,
+          "brier": 0.2647027577519989,
+          "ece": 0.009998726121221617,
+          "mean_conf": 0.7984406564077068,
+          "score_mae": 0.3859622411504242,
+          "cov@0.5": 0.9168552036199095,
+          "acc@0.5": 0.8339502364795394,
+          "cov@0.7": 0.7034313725490197,
+          "acc@0.7": 0.908201554543018,
+          "cov@0.9": 0.4518288084464555,
+          "acc@0.9": 0.9670352597538077
+        },
+        "tasks": [
+          "ag_news",
+          "banking77",
+          "boolq",
+          "clinc150",
+          "commonsense_qa",
+          "copa",
+          "dbpedia",
+          "emotion",
+          "facts",
+          "fits",
+          "imdb",
+          "jailbreak",
+          "massive_intent",
+          "mnli",
+          "mrpc",
+          "openbookqa",
+          "paws",
+          "qnli",
+          "read",
+          "rte",
+          "sciq",
+          "sst2",
+          "swag",
+          "tweet_emoji",
+          "tweet_hate",
+          "tweet_irony",
+          "tweet_offensive",
+          "tweet_sentiment",
+          "yahoo",
+          "yelp",
+          "anli",
+          "winogrande",
+          "hellaswag",
+          "race",
+          "scitail",
+          "qqp",
+          "stsb",
+          "toxic",
+          "stance_abortion",
+          "stance_atheism",
+          "stance_feminist",
+          "stance_hillary",
+          "match",
+          "goemo_soft",
+          "reason",
+          "formality",
+          "politeness",
+          "strategyqa",
+          "vitaminc",
+          "ruletaker",
+          "proofwriter",
+          "folio",
+          "logiqa",
+          "tracie",
+          "temporal_nli",
+          "piqa",
+          "siqa",
+          "clutrr",
+          "gsm8k",
+          "svamp",
+          "aqua"
+        ]
+      }
+    },
+    "v0007-spam": {
+      "version": "v0007-spam",
+      "parent": "v0007",
+      "parents": [
+        "v0007"
+      ],
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-18T06:05:30+00:00",
+      "tasks": [
+        "spam"
+      ],
+      "trained_on": [
+        "spam"
+      ],
+      "holdout": [
+        "probe",
+        "sst2",
+        "mnli",
+        "cola"
+      ],
+      "steps": 589,
+      "train_examples": 4000,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 8,
+        "loss": "mix",
+        "max_per_task": 4000,
+        "epochs": 1,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {
+        "spam": {
+          "n": 300,
+          "acc": 0.99,
+          "nll": 0.04094129466820647,
+          "brier": 0.01309215253075488,
+          "ece": 0.031613957881927515,
+          "mean_conf": 0.9694881041844686,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.99,
+          "cov@0.7": 0.9933333333333333,
+          "acc@0.7": 0.9966442953020134,
+          "cov@0.9": 0.9666666666666667,
+          "acc@0.9": 1.0
+        }
+      },
+      "unseen_test_uncalibrated": {
+        "probe": {
+          "n": 97,
+          "acc": 0.8969072164948454,
+          "nll": 0.21773648728926945,
+          "brier": 0.1273214505148998,
+          "ece": 0.07348626329726782,
+          "mean_conf": 0.9161194676590949,
+          "score_mae": 0.15704563955659978,
+          "cov@0.5": 0.9896907216494846,
+          "acc@0.5": 0.90625,
+          "cov@0.7": 0.8969072164948454,
+          "acc@0.7": 0.9540229885057471,
+          "cov@0.9": 0.7938144329896907,
+          "acc@0.9": 0.987012987012987,
+          "families": {
+            "desc": [
+              13,
+              15
+            ],
+            "negation": [
+              10,
+              10
+            ],
+            "logic": [
+              13,
+              17
+            ],
+            "score": [
+              11,
+              12
+            ],
+            "json": [
+              6,
+              6
+            ],
+            "taxonomy": [
+              14,
+              14
+            ],
+            "plausible": [
+              4,
+              4
+            ],
+            "twist": [
+              3,
+              3
+            ],
+            "time": [
+              4,
+              5
+            ],
+            "intent": [
+              2,
+              3
+            ],
+            "compare": [
+              3,
+              4
+            ],
+            "criteria": [
+              4,
+              4
+            ]
+          }
+        },
+        "sst2": {
+          "n": 872,
+          "acc": 0.9575688073394495,
+          "nll": 0.1396011574813463,
+          "brier": 0.07172239838956931,
+          "ece": 0.014292597087151384,
+          "mean_conf": 0.9680821417121712,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.9575688073394495,
+          "cov@0.7": 0.9793577981651376,
+          "acc@0.7": 0.9637002341920374,
+          "cov@0.9": 0.9438073394495413,
+          "acc@0.9": 0.9732685297691372
+        },
+        "mnli": {
+          "n": 1000,
+          "acc": 0.859,
+          "nll": 0.3888388354725572,
+          "brier": 0.21223603757591059,
+          "ece": 0.04788112017512322,
+          "mean_conf": 0.8856889481842518,
+          "cov@0.5": 0.987,
+          "acc@0.5": 0.8662613981762918,
+          "cov@0.7": 0.909,
+          "acc@0.7": 0.8954895489548955,
+          "cov@0.9": 0.655,
+          "acc@0.9": 0.9587786259541985
+        },
+        "cola": {
+          "n": 1000,
+          "acc": 0.759,
+          "nll": 0.5066963657737495,
+          "brier": 0.33443985155856215,
+          "ece": 0.0523103475570679,
+          "mean_conf": 0.791339822769165,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.759,
+          "cov@0.7": 0.764,
+          "acc@0.7": 0.8167539267015707,
+          "cov@0.9": 0.207,
+          "acc@0.9": 0.9516908212560387
+        }
+      },
+      "summary": {
+        "mean_acc": 0.99,
+        "mean_ece": 0.031613957881927515
+      },
+      "history": []
+    },
+    "v0007-support": {
+      "version": "v0007-support",
+      "parent": "v0007",
+      "parents": [
+        "v0007"
+      ],
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-18T12:52:37+00:00",
+      "tasks": [
+        "banking77",
+        "clinc150",
+        "massive_intent"
+      ],
+      "trained_on": [
+        "banking77",
+        "clinc150",
+        "massive_intent"
+      ],
+      "holdout": [
+        "trec"
+      ],
+      "steps": 1067,
+      "train_examples": 12000,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 8,
+        "loss": "mix",
+        "max_per_task": 4000,
+        "epochs": 1,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {},
+      "unseen_test_uncalibrated": {
+        "trec": {
+          "n": 500,
+          "acc": 0.758,
+          "nll": 0.767611707548029,
+          "brier": 0.37113630112335455,
+          "ece": 0.08773053616285323,
+          "mean_conf": 0.8239709965586662,
+          "cov@0.5": 0.924,
+          "acc@0.5": 0.7835497835497836,
+          "cov@0.7": 0.762,
+          "acc@0.7": 0.8293963254593176,
+          "cov@0.9": 0.474,
+          "acc@0.9": 0.8734177215189873
+        }
+      },
+      "summary": {
+        "mean_acc": null,
+        "mean_ece": null
+      },
+      "history": []
+    },
+    "v0007-safety": {
+      "version": "v0007-safety",
+      "parent": "v0007",
+      "parents": [
+        "v0007"
+      ],
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-18T13:21:25+00:00",
+      "tasks": [
+        "jailbreak",
+        "toxic",
+        "tweet_hate",
+        "tweet_offensive"
+      ],
+      "trained_on": [
+        "jailbreak",
+        "toxic",
+        "tweet_hate",
+        "tweet_offensive"
+      ],
+      "holdout": [],
+      "steps": 607,
+      "train_examples": 8344,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 8,
+        "loss": "mix",
+        "max_per_task": 2500,
+        "epochs": 1,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {
+        "jailbreak": {
+          "n": 200,
+          "acc": 0.98,
+          "nll": 0.06724073947718683,
+          "brier": 0.03142695878535079,
+          "ece": 0.008043854534626017,
+          "mean_conf": 0.988043854534626,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.98,
+          "cov@0.7": 0.995,
+          "acc@0.7": 0.9849246231155779,
+          "cov@0.9": 0.99,
+          "acc@0.9": 0.98989898989899
+        },
+        "toxic": {
+          "n": 300,
+          "acc": 0.8666666666666667,
+          "nll": 0.3084903371258136,
+          "brier": 0.19258679203988513,
+          "ece": 0.03767327169577279,
+          "mean_conf": 0.8952765788634618,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.8666666666666667,
+          "cov@0.7": 0.8966666666666666,
+          "acc@0.7": 0.9033457249070632,
+          "cov@0.9": 0.6766666666666666,
+          "acc@0.9": 0.9655172413793104
+        },
+        "tweet_hate": {
+          "n": 300,
+          "acc": 0.7433333333333333,
+          "nll": 0.5153975753819442,
+          "brier": 0.3425311312237358,
+          "ece": 0.08356260061264038,
+          "mean_conf": 0.8240628039836884,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.7433333333333333,
+          "cov@0.7": 0.78,
+          "acc@0.7": 0.811965811965812,
+          "cov@0.9": 0.39,
+          "acc@0.9": 0.9145299145299145
+        },
+        "tweet_offensive": {
+          "n": 300,
+          "acc": 0.7866666666666666,
+          "nll": 0.4662258180163463,
+          "brier": 0.3043113272015309,
+          "ece": 0.053681320548057555,
+          "mean_conf": 0.8303003575404485,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.7866666666666666,
+          "cov@0.7": 0.7933333333333333,
+          "acc@0.7": 0.8403361344537815,
+          "cov@0.9": 0.4533333333333333,
+          "acc@0.9": 0.9191176470588235
+        }
+      },
+      "unseen_test_uncalibrated": {},
+      "summary": {
+        "mean_acc": 0.8441666666666666,
+        "mean_ece": 0.04574026184777419
+      },
+      "history": []
+    },
+    "v0007-reason": {
+      "version": "v0007-reason",
+      "parent": "v0007",
+      "parents": [
+        "v0007"
+      ],
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-22T05:53:47+00:00",
+      "tasks": [
+        "traps",
+        "reason",
+        "skills"
+      ],
+      "trained_on": [
+        "traps",
+        "reason",
+        "skills"
+      ],
+      "holdout": [
+        "probe"
+      ],
+      "steps": 3144,
+      "train_examples": 49321,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 16,
+        "loss": "mix",
+        "max_per_task": 22000,
+        "epochs": 1,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {
+        "traps": {
+          "n": 300,
+          "acc": 0.9466666666666667,
+          "nll": 0.20538718146178303,
+          "brier": 0.06905301894228903,
+          "ece": 0.04965091367562613,
+          "mean_conf": 0.92676440179348,
+          "score_mae": 0.47862898526946085,
+          "cov@0.5": 0.9866666666666667,
+          "acc@0.5": 0.9493243243243243,
+          "cov@0.7": 0.9466666666666667,
+          "acc@0.7": 0.971830985915493,
+          "cov@0.9": 0.8433333333333334,
+          "acc@0.9": 0.9920948616600791
+        },
+        "reason": {
+          "n": 300,
+          "acc": 0.98,
+          "nll": 0.046410952195114846,
+          "brier": 0.025619819302402664,
+          "ece": 0.025865860184033737,
+          "mean_conf": 0.9718742754062016,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.98,
+          "cov@0.7": 0.97,
+          "acc@0.7": 0.9896907216494846,
+          "cov@0.9": 0.95,
+          "acc@0.9": 1.0
+        },
+        "skills": {
+          "n": 300,
+          "acc": 0.9033333333333333,
+          "nll": 0.20629042731885439,
+          "brier": 0.11973303077943658,
+          "ece": 0.061399669448534644,
+          "mean_conf": 0.8790679361422856,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.9033333333333333,
+          "cov@0.7": 0.83,
+          "acc@0.7": 0.9759036144578314,
+          "cov@0.9": 0.7166666666666667,
+          "acc@0.9": 1.0
+        }
+      },
+      "unseen_test_uncalibrated": {
+        "probe": {
+          "n": 97,
+          "acc": 0.9381443298969072,
+          "nll": 0.1893354459761264,
+          "brier": 0.09341768317288839,
+          "ece": 0.04105151006855912,
+          "mean_conf": 0.9446589375279614,
+          "score_mae": 0.18023038718577786,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.9381443298969072,
+          "cov@0.7": 0.9381443298969072,
+          "acc@0.7": 0.967032967032967,
+          "cov@0.9": 0.865979381443299,
+          "acc@0.9": 0.9761904761904762,
+          "families": {
+            "desc": [
+              14,
+              15
+            ],
+            "negation": [
+              10,
+              10
+            ],
+            "logic": [
+              15,
+              17
+            ],
+            "score": [
+              11,
+              12
+            ],
+            "json": [
+              6,
+              6
+            ],
+            "taxonomy": [
+              14,
+              14
+            ],
+            "plausible": [
+              4,
+              4
+            ],
+            "twist": [
+              3,
+              3
+            ],
+            "time": [
+              4,
+              5
+            ],
+            "intent": [
+              2,
+              3
+            ],
+            "compare": [
+              4,
+              4
+            ],
+            "criteria": [
+              4,
+              4
+            ]
+          }
+        }
+      },
+      "summary": {
+        "mean_acc": 0.9433333333333334,
+        "mean_ece": 0.04563881443606484
+      },
+      "history": []
+    },
+    "v0007-language": {
+      "version": "v0007-language",
+      "parent": "v0007",
+      "parents": [
+        "v0007"
+      ],
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-22T05:59:59+00:00",
+      "tasks": [
+        "language"
+      ],
+      "trained_on": [
+        "language"
+      ],
+      "holdout": [],
+      "steps": 625,
+      "train_examples": 10000,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 8,
+        "loss": "mix",
+        "max_per_task": 10000,
+        "epochs": 1,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {
+        "language": {
+          "n": 300,
+          "acc": 1.0,
+          "nll": 0.11204298576288962,
+          "brier": 0.0022332283333394826,
+          "ece": 0.05060336291790009,
+          "mean_conf": 0.9493966370820999,
+          "cov@0.5": 1.0,
+          "acc@0.5": 1.0,
+          "cov@0.7": 1.0,
+          "acc@0.7": 1.0,
+          "cov@0.9": 0.99,
+          "acc@0.9": 1.0
+        }
+      },
+      "unseen_test_uncalibrated": {},
+      "summary": {
+        "mean_acc": 1.0,
+        "mean_ece": 0.05060336291790009
+      },
+      "history": []
+    },
+    "v0007-sentiment": {
+      "version": "v0007-sentiment",
+      "parent": "v0007",
+      "parents": [
+        "v0007"
+      ],
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-22T06:27:25+00:00",
+      "tasks": [
+        "sst2",
+        "sst5",
+        "yelp",
+        "tweet_sentiment",
+        "emotion",
+        "goemo_soft",
+        "tweet_irony",
+        "imdb",
+        "formality",
+        "politeness",
+        "sarcasm"
+      ],
+      "trained_on": [
+        "sst2",
+        "sst5",
+        "yelp",
+        "tweet_sentiment",
+        "emotion",
+        "goemo_soft",
+        "tweet_irony",
+        "imdb",
+        "formality",
+        "politeness",
+        "sarcasm"
+      ],
+      "holdout": [
+        "fin_sentiment",
+        "counterfactual"
+      ],
+      "steps": 1782,
+      "train_examples": 22000,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 8,
+        "loss": "mix",
+        "max_per_task": 2000,
+        "epochs": 1,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {
+        "sst2": {
+          "n": 300,
+          "acc": 0.9466666666666667,
+          "nll": 0.1605975828277054,
+          "brier": 0.0888536530799026,
+          "ece": 0.027998056411743223,
+          "mean_conf": 0.9642687650521596,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.9466666666666667,
+          "cov@0.7": 0.9766666666666667,
+          "acc@0.7": 0.9556313993174061,
+          "cov@0.9": 0.9066666666666666,
+          "acc@0.9": 0.9632352941176471
+        },
+        "sst5": {
+          "n": 300,
+          "acc": 0.5766666666666667,
+          "nll": 1.030410210134918,
+          "brier": 0.5736773828010426,
+          "ece": 0.05870265493790308,
+          "mean_conf": 0.5829293202360472,
+          "score_mae": 0.5457962888351175,
+          "cov@0.5": 0.79,
+          "acc@0.5": 0.6075949367088608,
+          "cov@0.7": 0.11666666666666667,
+          "acc@0.7": 0.7142857142857143,
+          "cov@0.9": 0.0,
+          "acc@0.9": NaN
+        },
+        "yelp": {
+          "n": 300,
+          "acc": 0.69,
+          "nll": 0.7307452479655708,
+          "brier": 0.4235077727191108,
+          "ece": 0.08372766455014549,
+          "mean_conf": 0.7693662059307098,
+          "score_mae": 0.36723209824798686,
+          "cov@0.5": 0.9533333333333334,
+          "acc@0.5": 0.7027972027972028,
+          "cov@0.7": 0.69,
+          "acc@0.7": 0.782608695652174,
+          "cov@0.9": 0.21333333333333335,
+          "acc@0.9": 0.921875
+        },
+        "tweet_sentiment": {
+          "n": 300,
+          "acc": 0.7233333333333334,
+          "nll": 0.6324473977408053,
+          "brier": 0.37319421960896293,
+          "ece": 0.061867911020914726,
+          "mean_conf": 0.7788027099768321,
+          "score_mae": 0.3486348866733412,
+          "cov@0.5": 0.99,
+          "acc@0.5": 0.7239057239057239,
+          "cov@0.7": 0.69,
+          "acc@0.7": 0.8115942028985508,
+          "cov@0.9": 0.2733333333333333,
+          "acc@0.9": 0.9024390243902439
+        },
+        "emotion": {
+          "n": 300,
+          "acc": 0.78,
+          "nll": 0.6305815624709215,
+          "brier": 0.3066913792466958,
+          "ece": 0.07204169511795044,
+          "mean_conf": 0.8520416951179505,
+          "cov@0.5": 0.96,
+          "acc@0.5": 0.7951388888888888,
+          "cov@0.7": 0.8166666666666667,
+          "acc@0.7": 0.8612244897959184,
+          "cov@0.9": 0.55,
+          "acc@0.9": 0.9575757575757575
+        },
+        "tweet_irony": {
+          "n": 300,
+          "acc": 0.7366666666666667,
+          "nll": 0.5502091578371892,
+          "brier": 0.3718147638320287,
+          "ece": 0.08492387334505717,
+          "mean_conf": 0.7882708374659221,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.7366666666666667,
+          "cov@0.7": 0.74,
+          "acc@0.7": 0.7837837837837838,
+          "cov@0.9": 0.23333333333333334,
+          "acc@0.9": 0.9142857142857143
+        },
+        "imdb": {
+          "n": 300,
+          "acc": 0.9633333333333334,
+          "nll": 0.13351794521006682,
+          "brier": 0.060667604680717205,
+          "ece": 0.025028558770815598,
+          "mean_conf": 0.9825289577245713,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.9633333333333334,
+          "cov@0.7": 0.99,
+          "acc@0.7": 0.9696969696969697,
+          "cov@0.9": 0.9633333333333334,
+          "acc@0.9": 0.9792387543252595
+        },
+        "formality": {
+          "n": 300,
+          "acc": 0.6133333333333333,
+          "nll": 0.9954126704919857,
+          "brier": 0.25669049533824123,
+          "ece": 0.061814626157283774,
+          "mean_conf": 0.5946429490049681,
+          "score_mae": 0.4740935998460433,
+          "cov@0.5": 0.89,
+          "acc@0.5": 0.6179775280898876,
+          "cov@0.7": 0.06666666666666667,
+          "acc@0.7": 0.75,
+          "cov@0.9": 0.0,
+          "acc@0.9": NaN
+        },
+        "politeness": {
+          "n": 300,
+          "acc": 0.8566666666666667,
+          "nll": 0.3775658668930503,
+          "brier": 0.19322453000823264,
+          "ece": 0.03797433485587436,
+          "mean_conf": 0.8931960561871528,
+          "score_mae": 0.2005545631381392,
+          "cov@0.5": 0.9766666666666667,
+          "acc@0.5": 0.8703071672354948,
+          "cov@0.7": 0.87,
+          "acc@0.7": 0.9233716475095786,
+          "cov@0.9": 0.6633333333333333,
+          "acc@0.9": 0.9748743718592965
+        },
+        "sarcasm": {
+          "n": 300,
+          "acc": 0.8766666666666667,
+          "nll": 0.45201081925126335,
+          "brier": 0.10946457478491955,
+          "ece": 0.08731486479441326,
+          "mean_conf": 0.8211214067538579,
+          "score_mae": 0.4119935893134519,
+          "cov@0.5": 0.94,
+          "acc@0.5": 0.8971631205673759,
+          "cov@0.7": 0.7633333333333333,
+          "acc@0.7": 0.9606986899563319,
+          "cov@0.9": 0.49,
+          "acc@0.9": 0.9931972789115646
+        }
+      },
+      "unseen_test_uncalibrated": {
+        "fin_sentiment": {
+          "n": 1000,
+          "acc": 0.807,
+          "nll": 0.4482609479897583,
+          "brier": 0.2712111023947865,
+          "ece": 0.025465373486280456,
+          "mean_conf": 0.8211387673318387,
+          "cov@0.5": 0.992,
+          "acc@0.5": 0.8104838709677419,
+          "cov@0.7": 0.799,
+          "acc@0.7": 0.8685857321652065,
+          "cov@0.9": 0.369,
+          "acc@0.9": 0.9376693766937669
+        },
+        "counterfactual": {
+          "n": 1000,
+          "acc": 0.855,
+          "nll": 0.3709632075470155,
+          "brier": 0.2216712187729024,
+          "ece": 0.07320722198486329,
+          "mean_conf": 0.7820079625844956,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.855,
+          "cov@0.7": 0.748,
+          "acc@0.7": 0.9251336898395722,
+          "cov@0.9": 0.135,
+          "acc@0.9": 0.9703703703703703
+        }
+      },
+      "summary": {
+        "mean_acc": 0.7763333333333333,
+        "mean_ece": 0.06013942399621011
+      },
+      "history": []
+    },
+    "v0007-typed": {
+      "version": "v0007-typed",
+      "parent": "v0007",
+      "parents": [
+        "v0007"
+      ],
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-22T11:39:50+00:00",
+      "tasks": [
+        "typed_decisions"
+      ],
+      "trained_on": [
+        "typed_decisions"
+      ],
+      "holdout": [],
+      "steps": 1352,
+      "train_examples": 5400,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 16,
+        "loss": "mix",
+        "max_per_task": 6000,
+        "epochs": 2,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {
+        "typed_decisions": {
+          "n": 300,
+          "acc": 0.85,
+          "nll": 0.7897254475627804,
+          "brier": 0.04430931889174507,
+          "ece": 0.21966197381416955,
+          "mean_conf": 0.6374534544348717,
+          "score_mae": 0.34401526501434937,
+          "cov@0.5": 0.76,
+          "acc@0.5": 0.9342105263157895,
+          "cov@0.7": 0.30666666666666664,
+          "acc@0.7": 0.9891304347826086,
+          "cov@0.9": 0.13666666666666666,
+          "acc@0.9": 1.0
+        }
+      },
+      "unseen_test_uncalibrated": {},
+      "summary": {
+        "mean_acc": 0.85,
+        "mean_ece": 0.21966197381416955
+      },
+      "history": []
+    },
+    "v0007-math": {
+      "version": "v0007-math",
+      "parent": "v0007",
+      "parents": [
+        "v0007"
+      ],
+      "kind": "lmhead",
+      "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+      "max_len": 1024,
+      "created": "2026-09-22T13:55:28+00:00",
+      "tasks": [
+        "math",
+        "gsm8k",
+        "svamp",
+        "aqua"
+      ],
+      "trained_on": [
+        "math",
+        "gsm8k",
+        "svamp",
+        "aqua"
+      ],
+      "holdout": [],
+      "steps": 3684,
+      "train_examples": 57771,
+      "args": {
+        "cmd": "lmtrain",
+        "lora_r": 32,
+        "loss": "mix",
+        "max_per_task": 30000,
+        "epochs": 1,
+        "lr": 0.0001,
+        "anchor": 0.1
+      },
+      "metrics": {
+        "math": {
+          "n": 300,
+          "acc": 0.8733333333333333,
+          "nll": 0.2491237256136509,
+          "brier": 0.15517219901605317,
+          "ece": 0.04117123444875081,
+          "mean_conf": 0.8545713822046915,
+          "cov@0.5": 1.0,
+          "acc@0.5": 0.8733333333333333,
+          "cov@0.7": 0.7533333333333333,
+          "acc@0.7": 0.9690265486725663,
+          "cov@0.9": 0.7066666666666667,
+          "acc@0.9": 1.0
+        },
+        "gsm8k": {
+          "n": 300,
+          "acc": 0.8166666666666667,
+          "nll": 0.4634518215859158,
+          "brier": 0.24914535577158609,
+          "ece": 0.08174108674128854,
+          "mean_conf": 0.7950625959038734,
+          "cov@0.5": 0.9,
+          "acc@0.5": 0.8666666666666667,
+          "cov@0.7": 0.7,
+          "acc@0.7": 0.9238095238095239,
+          "cov@0.9": 0.4266666666666667,
+          "acc@0.9": 0.9921875
+        },
+        "svamp": {
+          "n": 100,
+          "acc": 0.68,
+          "nll": 0.7193946922798429,
+          "brier": 0.41086004277348953,
+          "ece": 0.07071352988481525,
+          "mean_conf": 0.7261623141169548,
+          "cov@0.5": 0.83,
+          "acc@0.5": 0.7349397590361446,
+          "cov@0.7": 0.58,
+          "acc@0.7": 0.8275862068965517,
+          "cov@0.9": 0.2,
+          "acc@0.9": 1.0
+        },
+        "aqua": {
+          "n": 253,
+          "acc": 0.391304347826087,
+          "nll": 1.414271918680051,
+          "brier": 0.7115495715758425,
+          "ece": 0.0654898358899143,
+          "mean_conf": 0.3998582398467384,
+          "cov@0.5": 0.20553359683794467,
+          "acc@0.5": 0.6538461538461539,
+          "cov@0.7": 0.06324110671936758,
+          "acc@0.7": 0.8125,
+          "cov@0.9": 0.007905138339920948,
+          "acc@0.9": 1.0
+        }
+      },
+      "unseen_test_uncalibrated": {},
+      "summary": {
+        "mean_acc": 0.6903260869565218,
+        "mean_ece": 0.06477892174119222
+      },
+      "history": []
+    }
+  }
+}
\ No newline at end of file
diff --git a/v0007-language/adapter/README.md b/v0007-language/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007-language/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007-language/adapter/adapter_config.json b/v0007-language/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e476e709a3e5081afbd4873695e4de389a2cf7fa
--- /dev/null
+++ b/v0007-language/adapter/adapter_config.json
@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "k_proj",
+    "q_proj",
+    "v_proj",
+    "o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007-language/adapter/adapter_model.safetensors b/v0007-language/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..892b13996ef0b77c9ad3a4a9542f1798bc817a80
--- /dev/null
+++ b/v0007-language/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6529d354a16a5a1adc4edfa6e448a23251d6acfc2795ccdcd0aedd4b9f68e0a4
+size 8745704
diff --git a/v0007-language/config.json b/v0007-language/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5a06253609d2e12a0d57418deca14a0d40ad552
--- /dev/null
+++ b/v0007-language/config.json
@@ -0,0 +1,11 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "parents": [
+    "v0007"
+  ],
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0
+}
\ No newline at end of file
diff --git a/v0007-language/manifest.json b/v0007-language/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..ba5f953d8cb0fbf328dda4f24eb8a2b894bb4528
--- /dev/null
+++ b/v0007-language/manifest.json
@@ -0,0 +1,51 @@
+{
+  "version": "v0007-language",
+  "parent": "v0007",
+  "parents": [
+    "v0007"
+  ],
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-22T05:59:59+00:00",
+  "tasks": [
+    "language"
+  ],
+  "trained_on": [
+    "language"
+  ],
+  "holdout": [],
+  "steps": 625,
+  "train_examples": 10000,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 8,
+    "loss": "mix",
+    "max_per_task": 10000,
+    "epochs": 1,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {
+    "language": {
+      "n": 300,
+      "acc": 1.0,
+      "nll": 0.11204298576288962,
+      "brier": 0.0022332283333394826,
+      "ece": 0.05060336291790009,
+      "mean_conf": 0.9493966370820999,
+      "cov@0.5": 1.0,
+      "acc@0.5": 1.0,
+      "cov@0.7": 1.0,
+      "acc@0.7": 1.0,
+      "cov@0.9": 0.99,
+      "acc@0.9": 1.0
+    }
+  },
+  "unseen_test_uncalibrated": {},
+  "summary": {
+    "mean_acc": 1.0,
+    "mean_ece": 0.05060336291790009
+  },
+  "history": []
+}
\ No newline at end of file
diff --git a/v0007-math/adapter/README.md b/v0007-math/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007-math/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007-math/adapter/adapter_config.json b/v0007-math/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..4f157b7e488347e8d3ade79228745a72c00e0b78
--- /dev/null
+++ b/v0007-math/adapter/adapter_config.json
@@ -0,0 +1,51 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "k_proj",
+    "o_proj",
+    "v_proj",
+    "up_proj",
+    "q_proj",
+    "down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007-math/adapter/adapter_model.safetensors b/v0007-math/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..e7fedcc4edc0be7bea591b34f8b1f380da7f1882
--- /dev/null
+++ b/v0007-math/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2093256ec28e1fe21689add9ea3c8434bfac3ba07c77644505ad375a7b7e4c9b
+size 147770496
diff --git a/v0007-math/config.json b/v0007-math/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5a06253609d2e12a0d57418deca14a0d40ad552
--- /dev/null
+++ b/v0007-math/config.json
@@ -0,0 +1,11 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "parents": [
+    "v0007"
+  ],
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0
+}
\ No newline at end of file
diff --git a/v0007-math/manifest.json b/v0007-math/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..9ccf950e382b0543e8246a15dfa49999ae579f5b
--- /dev/null
+++ b/v0007-math/manifest.json
@@ -0,0 +1,99 @@
+{
+  "version": "v0007-math",
+  "parent": "v0007",
+  "parents": [
+    "v0007"
+  ],
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-22T13:55:28+00:00",
+  "tasks": [
+    "math",
+    "gsm8k",
+    "svamp",
+    "aqua"
+  ],
+  "trained_on": [
+    "math",
+    "gsm8k",
+    "svamp",
+    "aqua"
+  ],
+  "holdout": [],
+  "steps": 3684,
+  "train_examples": 57771,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 32,
+    "loss": "mix",
+    "max_per_task": 30000,
+    "epochs": 1,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {
+    "math": {
+      "n": 300,
+      "acc": 0.8733333333333333,
+      "nll": 0.2491237256136509,
+      "brier": 0.15517219901605317,
+      "ece": 0.04117123444875081,
+      "mean_conf": 0.8545713822046915,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.8733333333333333,
+      "cov@0.7": 0.7533333333333333,
+      "acc@0.7": 0.9690265486725663,
+      "cov@0.9": 0.7066666666666667,
+      "acc@0.9": 1.0
+    },
+    "gsm8k": {
+      "n": 300,
+      "acc": 0.8166666666666667,
+      "nll": 0.4634518215859158,
+      "brier": 0.24914535577158609,
+      "ece": 0.08174108674128854,
+      "mean_conf": 0.7950625959038734,
+      "cov@0.5": 0.9,
+      "acc@0.5": 0.8666666666666667,
+      "cov@0.7": 0.7,
+      "acc@0.7": 0.9238095238095239,
+      "cov@0.9": 0.4266666666666667,
+      "acc@0.9": 0.9921875
+    },
+    "svamp": {
+      "n": 100,
+      "acc": 0.68,
+      "nll": 0.7193946922798429,
+      "brier": 0.41086004277348953,
+      "ece": 0.07071352988481525,
+      "mean_conf": 0.7261623141169548,
+      "cov@0.5": 0.83,
+      "acc@0.5": 0.7349397590361446,
+      "cov@0.7": 0.58,
+      "acc@0.7": 0.8275862068965517,
+      "cov@0.9": 0.2,
+      "acc@0.9": 1.0
+    },
+    "aqua": {
+      "n": 253,
+      "acc": 0.391304347826087,
+      "nll": 1.414271918680051,
+      "brier": 0.7115495715758425,
+      "ece": 0.0654898358899143,
+      "mean_conf": 0.3998582398467384,
+      "cov@0.5": 0.20553359683794467,
+      "acc@0.5": 0.6538461538461539,
+      "cov@0.7": 0.06324110671936758,
+      "acc@0.7": 0.8125,
+      "cov@0.9": 0.007905138339920948,
+      "acc@0.9": 1.0
+    }
+  },
+  "unseen_test_uncalibrated": {},
+  "summary": {
+    "mean_acc": 0.6903260869565218,
+    "mean_ece": 0.06477892174119222
+  },
+  "history": []
+}
\ No newline at end of file
diff --git a/v0007-reason/adapter/README.md b/v0007-reason/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007-reason/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007-reason/adapter/adapter_config.json b/v0007-reason/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..d8e1677fe12b54115962c22a3b22e7eef353a5e4
--- /dev/null
+++ b/v0007-reason/adapter/adapter_config.json
@@ -0,0 +1,51 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "up_proj",
+    "q_proj",
+    "o_proj",
+    "down_proj",
+    "gate_proj",
+    "k_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007-reason/adapter/adapter_model.safetensors b/v0007-reason/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..c886fea3ff4e3c21cfa2458db6c75517676d9576
--- /dev/null
+++ b/v0007-reason/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fc7c0ac7ff588b8265990a7b7fc80d48483dec2e776c09d3ddb8c6b6146b0f0
+size 73911112
diff --git a/v0007-reason/config.json b/v0007-reason/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5a06253609d2e12a0d57418deca14a0d40ad552
--- /dev/null
+++ b/v0007-reason/config.json
@@ -0,0 +1,11 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "parents": [
+    "v0007"
+  ],
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0
+}
\ No newline at end of file
diff --git a/v0007-reason/manifest.json b/v0007-reason/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..5822777b0949b5b48a33f2571de9091c15dafb21
--- /dev/null
+++ b/v0007-reason/manifest.json
@@ -0,0 +1,152 @@
+{
+  "version": "v0007-reason",
+  "parent": "v0007",
+  "parents": [
+    "v0007"
+  ],
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-22T05:53:47+00:00",
+  "tasks": [
+    "traps",
+    "reason",
+    "skills"
+  ],
+  "trained_on": [
+    "traps",
+    "reason",
+    "skills"
+  ],
+  "holdout": [
+    "probe"
+  ],
+  "steps": 3144,
+  "train_examples": 49321,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 16,
+    "loss": "mix",
+    "max_per_task": 22000,
+    "epochs": 1,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {
+    "traps": {
+      "n": 300,
+      "acc": 0.9466666666666667,
+      "nll": 0.20538718146178303,
+      "brier": 0.06905301894228903,
+      "ece": 0.04965091367562613,
+      "mean_conf": 0.92676440179348,
+      "score_mae": 0.47862898526946085,
+      "cov@0.5": 0.9866666666666667,
+      "acc@0.5": 0.9493243243243243,
+      "cov@0.7": 0.9466666666666667,
+      "acc@0.7": 0.971830985915493,
+      "cov@0.9": 0.8433333333333334,
+      "acc@0.9": 0.9920948616600791
+    },
+    "reason": {
+      "n": 300,
+      "acc": 0.98,
+      "nll": 0.046410952195114846,
+      "brier": 0.025619819302402664,
+      "ece": 0.025865860184033737,
+      "mean_conf": 0.9718742754062016,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.98,
+      "cov@0.7": 0.97,
+      "acc@0.7": 0.9896907216494846,
+      "cov@0.9": 0.95,
+      "acc@0.9": 1.0
+    },
+    "skills": {
+      "n": 300,
+      "acc": 0.9033333333333333,
+      "nll": 0.20629042731885439,
+      "brier": 0.11973303077943658,
+      "ece": 0.061399669448534644,
+      "mean_conf": 0.8790679361422856,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.9033333333333333,
+      "cov@0.7": 0.83,
+      "acc@0.7": 0.9759036144578314,
+      "cov@0.9": 0.7166666666666667,
+      "acc@0.9": 1.0
+    }
+  },
+  "unseen_test_uncalibrated": {
+    "probe": {
+      "n": 97,
+      "acc": 0.9381443298969072,
+      "nll": 0.1893354459761264,
+      "brier": 0.09341768317288839,
+      "ece": 0.04105151006855912,
+      "mean_conf": 0.9446589375279614,
+      "score_mae": 0.18023038718577786,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.9381443298969072,
+      "cov@0.7": 0.9381443298969072,
+      "acc@0.7": 0.967032967032967,
+      "cov@0.9": 0.865979381443299,
+      "acc@0.9": 0.9761904761904762,
+      "families": {
+        "desc": [
+          14,
+          15
+        ],
+        "negation": [
+          10,
+          10
+        ],
+        "logic": [
+          15,
+          17
+        ],
+        "score": [
+          11,
+          12
+        ],
+        "json": [
+          6,
+          6
+        ],
+        "taxonomy": [
+          14,
+          14
+        ],
+        "plausible": [
+          4,
+          4
+        ],
+        "twist": [
+          3,
+          3
+        ],
+        "time": [
+          4,
+          5
+        ],
+        "intent": [
+          2,
+          3
+        ],
+        "compare": [
+          4,
+          4
+        ],
+        "criteria": [
+          4,
+          4
+        ]
+      }
+    }
+  },
+  "summary": {
+    "mean_acc": 0.9433333333333334,
+    "mean_ece": 0.04563881443606484
+  },
+  "history": []
+}
\ No newline at end of file
diff --git a/v0007-safety/adapter/README.md b/v0007-safety/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007-safety/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007-safety/adapter/adapter_config.json b/v0007-safety/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..6d15eacc534b9c1dd8c8da260b86901dd63425ee
--- /dev/null
+++ b/v0007-safety/adapter/adapter_config.json
@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "k_proj",
+    "o_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007-safety/adapter/adapter_model.safetensors b/v0007-safety/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..3564ef178b16f9508caf10d8b9babf4e6e32a0ef
--- /dev/null
+++ b/v0007-safety/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5f965c1e859e8aa35ef4f37b347ca108e50f84b7e8cce29eee581ffe48f2de0
+size 8745704
diff --git a/v0007-safety/config.json b/v0007-safety/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5a06253609d2e12a0d57418deca14a0d40ad552
--- /dev/null
+++ b/v0007-safety/config.json
@@ -0,0 +1,11 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "parents": [
+    "v0007"
+  ],
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0
+}
\ No newline at end of file
diff --git a/v0007-safety/manifest.json b/v0007-safety/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..37bd45532865bd249826cd763a3677c862a443e4
--- /dev/null
+++ b/v0007-safety/manifest.json
@@ -0,0 +1,99 @@
+{
+  "version": "v0007-safety",
+  "parent": "v0007",
+  "parents": [
+    "v0007"
+  ],
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-18T13:21:25+00:00",
+  "tasks": [
+    "jailbreak",
+    "toxic",
+    "tweet_hate",
+    "tweet_offensive"
+  ],
+  "trained_on": [
+    "jailbreak",
+    "toxic",
+    "tweet_hate",
+    "tweet_offensive"
+  ],
+  "holdout": [],
+  "steps": 607,
+  "train_examples": 8344,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 8,
+    "loss": "mix",
+    "max_per_task": 2500,
+    "epochs": 1,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {
+    "jailbreak": {
+      "n": 200,
+      "acc": 0.98,
+      "nll": 0.06724073947718683,
+      "brier": 0.03142695878535079,
+      "ece": 0.008043854534626017,
+      "mean_conf": 0.988043854534626,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.98,
+      "cov@0.7": 0.995,
+      "acc@0.7": 0.9849246231155779,
+      "cov@0.9": 0.99,
+      "acc@0.9": 0.98989898989899
+    },
+    "toxic": {
+      "n": 300,
+      "acc": 0.8666666666666667,
+      "nll": 0.3084903371258136,
+      "brier": 0.19258679203988513,
+      "ece": 0.03767327169577279,
+      "mean_conf": 0.8952765788634618,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.8666666666666667,
+      "cov@0.7": 0.8966666666666666,
+      "acc@0.7": 0.9033457249070632,
+      "cov@0.9": 0.6766666666666666,
+      "acc@0.9": 0.9655172413793104
+    },
+    "tweet_hate": {
+      "n": 300,
+      "acc": 0.7433333333333333,
+      "nll": 0.5153975753819442,
+      "brier": 0.3425311312237358,
+      "ece": 0.08356260061264038,
+      "mean_conf": 0.8240628039836884,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.7433333333333333,
+      "cov@0.7": 0.78,
+      "acc@0.7": 0.811965811965812,
+      "cov@0.9": 0.39,
+      "acc@0.9": 0.9145299145299145
+    },
+    "tweet_offensive": {
+      "n": 300,
+      "acc": 0.7866666666666666,
+      "nll": 0.4662258180163463,
+      "brier": 0.3043113272015309,
+      "ece": 0.053681320548057555,
+      "mean_conf": 0.8303003575404485,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.7866666666666666,
+      "cov@0.7": 0.7933333333333333,
+      "acc@0.7": 0.8403361344537815,
+      "cov@0.9": 0.4533333333333333,
+      "acc@0.9": 0.9191176470588235
+    }
+  },
+  "unseen_test_uncalibrated": {},
+  "summary": {
+    "mean_acc": 0.8441666666666666,
+    "mean_ece": 0.04574026184777419
+  },
+  "history": []
+}
\ No newline at end of file
diff --git a/v0007-sentiment/adapter/README.md b/v0007-sentiment/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007-sentiment/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007-sentiment/adapter/adapter_config.json b/v0007-sentiment/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..91f44ee2ab6f2463328be7e44c953e19987582b0
--- /dev/null
+++ b/v0007-sentiment/adapter/adapter_config.json
@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "o_proj",
+    "v_proj",
+    "q_proj",
+    "k_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007-sentiment/adapter/adapter_model.safetensors b/v0007-sentiment/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..ce8a36460f9a383fecce08fd06446268b9a706eb
--- /dev/null
+++ b/v0007-sentiment/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3852848d0d44b8247bae59335f4a5a8a8a01cd798a324471378a451ceb8a3e6e
+size 8745704
diff --git a/v0007-sentiment/config.json b/v0007-sentiment/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5a06253609d2e12a0d57418deca14a0d40ad552
--- /dev/null
+++ b/v0007-sentiment/config.json
@@ -0,0 +1,11 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "parents": [
+    "v0007"
+  ],
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0
+}
\ No newline at end of file
diff --git a/v0007-sentiment/manifest.json b/v0007-sentiment/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..d6031731220a2e4b823552e381edfa5cec94a823
--- /dev/null
+++ b/v0007-sentiment/manifest.json
@@ -0,0 +1,235 @@
+{
+  "version": "v0007-sentiment",
+  "parent": "v0007",
+  "parents": [
+    "v0007"
+  ],
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-22T06:27:25+00:00",
+  "tasks": [
+    "sst2",
+    "sst5",
+    "yelp",
+    "tweet_sentiment",
+    "emotion",
+    "goemo_soft",
+    "tweet_irony",
+    "imdb",
+    "formality",
+    "politeness",
+    "sarcasm"
+  ],
+  "trained_on": [
+    "sst2",
+    "sst5",
+    "yelp",
+    "tweet_sentiment",
+    "emotion",
+    "goemo_soft",
+    "tweet_irony",
+    "imdb",
+    "formality",
+    "politeness",
+    "sarcasm"
+  ],
+  "holdout": [
+    "fin_sentiment",
+    "counterfactual"
+  ],
+  "steps": 1782,
+  "train_examples": 22000,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 8,
+    "loss": "mix",
+    "max_per_task": 2000,
+    "epochs": 1,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {
+    "sst2": {
+      "n": 300,
+      "acc": 0.9466666666666667,
+      "nll": 0.1605975828277054,
+      "brier": 0.0888536530799026,
+      "ece": 0.027998056411743223,
+      "mean_conf": 0.9642687650521596,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.9466666666666667,
+      "cov@0.7": 0.9766666666666667,
+      "acc@0.7": 0.9556313993174061,
+      "cov@0.9": 0.9066666666666666,
+      "acc@0.9": 0.9632352941176471
+    },
+    "sst5": {
+      "n": 300,
+      "acc": 0.5766666666666667,
+      "nll": 1.030410210134918,
+      "brier": 0.5736773828010426,
+      "ece": 0.05870265493790308,
+      "mean_conf": 0.5829293202360472,
+      "score_mae": 0.5457962888351175,
+      "cov@0.5": 0.79,
+      "acc@0.5": 0.6075949367088608,
+      "cov@0.7": 0.11666666666666667,
+      "acc@0.7": 0.7142857142857143,
+      "cov@0.9": 0.0,
+      "acc@0.9": NaN
+    },
+    "yelp": {
+      "n": 300,
+      "acc": 0.69,
+      "nll": 0.7307452479655708,
+      "brier": 0.4235077727191108,
+      "ece": 0.08372766455014549,
+      "mean_conf": 0.7693662059307098,
+      "score_mae": 0.36723209824798686,
+      "cov@0.5": 0.9533333333333334,
+      "acc@0.5": 0.7027972027972028,
+      "cov@0.7": 0.69,
+      "acc@0.7": 0.782608695652174,
+      "cov@0.9": 0.21333333333333335,
+      "acc@0.9": 0.921875
+    },
+    "tweet_sentiment": {
+      "n": 300,
+      "acc": 0.7233333333333334,
+      "nll": 0.6324473977408053,
+      "brier": 0.37319421960896293,
+      "ece": 0.061867911020914726,
+      "mean_conf": 0.7788027099768321,
+      "score_mae": 0.3486348866733412,
+      "cov@0.5": 0.99,
+      "acc@0.5": 0.7239057239057239,
+      "cov@0.7": 0.69,
+      "acc@0.7": 0.8115942028985508,
+      "cov@0.9": 0.2733333333333333,
+      "acc@0.9": 0.9024390243902439
+    },
+    "emotion": {
+      "n": 300,
+      "acc": 0.78,
+      "nll": 0.6305815624709215,
+      "brier": 0.3066913792466958,
+      "ece": 0.07204169511795044,
+      "mean_conf": 0.8520416951179505,
+      "cov@0.5": 0.96,
+      "acc@0.5": 0.7951388888888888,
+      "cov@0.7": 0.8166666666666667,
+      "acc@0.7": 0.8612244897959184,
+      "cov@0.9": 0.55,
+      "acc@0.9": 0.9575757575757575
+    },
+    "tweet_irony": {
+      "n": 300,
+      "acc": 0.7366666666666667,
+      "nll": 0.5502091578371892,
+      "brier": 0.3718147638320287,
+      "ece": 0.08492387334505717,
+      "mean_conf": 0.7882708374659221,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.7366666666666667,
+      "cov@0.7": 0.74,
+      "acc@0.7": 0.7837837837837838,
+      "cov@0.9": 0.23333333333333334,
+      "acc@0.9": 0.9142857142857143
+    },
+    "imdb": {
+      "n": 300,
+      "acc": 0.9633333333333334,
+      "nll": 0.13351794521006682,
+      "brier": 0.060667604680717205,
+      "ece": 0.025028558770815598,
+      "mean_conf": 0.9825289577245713,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.9633333333333334,
+      "cov@0.7": 0.99,
+      "acc@0.7": 0.9696969696969697,
+      "cov@0.9": 0.9633333333333334,
+      "acc@0.9": 0.9792387543252595
+    },
+    "formality": {
+      "n": 300,
+      "acc": 0.6133333333333333,
+      "nll": 0.9954126704919857,
+      "brier": 0.25669049533824123,
+      "ece": 0.061814626157283774,
+      "mean_conf": 0.5946429490049681,
+      "score_mae": 0.4740935998460433,
+      "cov@0.5": 0.89,
+      "acc@0.5": 0.6179775280898876,
+      "cov@0.7": 0.06666666666666667,
+      "acc@0.7": 0.75,
+      "cov@0.9": 0.0,
+      "acc@0.9": NaN
+    },
+    "politeness": {
+      "n": 300,
+      "acc": 0.8566666666666667,
+      "nll": 0.3775658668930503,
+      "brier": 0.19322453000823264,
+      "ece": 0.03797433485587436,
+      "mean_conf": 0.8931960561871528,
+      "score_mae": 0.2005545631381392,
+      "cov@0.5": 0.9766666666666667,
+      "acc@0.5": 0.8703071672354948,
+      "cov@0.7": 0.87,
+      "acc@0.7": 0.9233716475095786,
+      "cov@0.9": 0.6633333333333333,
+      "acc@0.9": 0.9748743718592965
+    },
+    "sarcasm": {
+      "n": 300,
+      "acc": 0.8766666666666667,
+      "nll": 0.45201081925126335,
+      "brier": 0.10946457478491955,
+      "ece": 0.08731486479441326,
+      "mean_conf": 0.8211214067538579,
+      "score_mae": 0.4119935893134519,
+      "cov@0.5": 0.94,
+      "acc@0.5": 0.8971631205673759,
+      "cov@0.7": 0.7633333333333333,
+      "acc@0.7": 0.9606986899563319,
+      "cov@0.9": 0.49,
+      "acc@0.9": 0.9931972789115646
+    }
+  },
+  "unseen_test_uncalibrated": {
+    "fin_sentiment": {
+      "n": 1000,
+      "acc": 0.807,
+      "nll": 0.4482609479897583,
+      "brier": 0.2712111023947865,
+      "ece": 0.025465373486280456,
+      "mean_conf": 0.8211387673318387,
+      "cov@0.5": 0.992,
+      "acc@0.5": 0.8104838709677419,
+      "cov@0.7": 0.799,
+      "acc@0.7": 0.8685857321652065,
+      "cov@0.9": 0.369,
+      "acc@0.9": 0.9376693766937669
+    },
+    "counterfactual": {
+      "n": 1000,
+      "acc": 0.855,
+      "nll": 0.3709632075470155,
+      "brier": 0.2216712187729024,
+      "ece": 0.07320722198486329,
+      "mean_conf": 0.7820079625844956,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.855,
+      "cov@0.7": 0.748,
+      "acc@0.7": 0.9251336898395722,
+      "cov@0.9": 0.135,
+      "acc@0.9": 0.9703703703703703
+    }
+  },
+  "summary": {
+    "mean_acc": 0.7763333333333333,
+    "mean_ece": 0.06013942399621011
+  },
+  "history": []
+}
\ No newline at end of file
diff --git a/v0007-spam/adapter/README.md b/v0007-spam/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007-spam/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007-spam/adapter/adapter_config.json b/v0007-spam/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..2bc4883c1f285f61a1160656078e601c30769451
--- /dev/null
+++ b/v0007-spam/adapter/adapter_config.json
@@ -0,0 +1,51 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "o_proj",
+    "up_proj",
+    "k_proj",
+    "gate_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007-spam/adapter/adapter_model.safetensors b/v0007-spam/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..bf93d5d4f57376b6afef74e86b016b1baf488e98
--- /dev/null
+++ b/v0007-spam/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2cdf75e1415b9c0771726ce43a0c4012dcb48bc9c1de8b68b7b718815312c7fb
+size 36981072
diff --git a/v0007-spam/config.json b/v0007-spam/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5a06253609d2e12a0d57418deca14a0d40ad552
--- /dev/null
+++ b/v0007-spam/config.json
@@ -0,0 +1,11 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "parents": [
+    "v0007"
+  ],
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0
+}
\ No newline at end of file
diff --git a/v0007-spam/manifest.json b/v0007-spam/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..e80e238251e986f9dd3ccf571d17b2e88b66ec6f
--- /dev/null
+++ b/v0007-spam/manifest.json
@@ -0,0 +1,164 @@
+{
+  "version": "v0007-spam",
+  "parent": "v0007",
+  "parents": [
+    "v0007"
+  ],
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-18T06:05:30+00:00",
+  "tasks": [
+    "spam"
+  ],
+  "trained_on": [
+    "spam"
+  ],
+  "holdout": [
+    "probe",
+    "sst2",
+    "mnli",
+    "cola"
+  ],
+  "steps": 589,
+  "train_examples": 4000,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 8,
+    "loss": "mix",
+    "max_per_task": 4000,
+    "epochs": 1,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {
+    "spam": {
+      "n": 300,
+      "acc": 0.99,
+      "nll": 0.04094129466820647,
+      "brier": 0.01309215253075488,
+      "ece": 0.031613957881927515,
+      "mean_conf": 0.9694881041844686,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.99,
+      "cov@0.7": 0.9933333333333333,
+      "acc@0.7": 0.9966442953020134,
+      "cov@0.9": 0.9666666666666667,
+      "acc@0.9": 1.0
+    }
+  },
+  "unseen_test_uncalibrated": {
+    "probe": {
+      "n": 97,
+      "acc": 0.8969072164948454,
+      "nll": 0.21773648728926945,
+      "brier": 0.1273214505148998,
+      "ece": 0.07348626329726782,
+      "mean_conf": 0.9161194676590949,
+      "score_mae": 0.15704563955659978,
+      "cov@0.5": 0.9896907216494846,
+      "acc@0.5": 0.90625,
+      "cov@0.7": 0.8969072164948454,
+      "acc@0.7": 0.9540229885057471,
+      "cov@0.9": 0.7938144329896907,
+      "acc@0.9": 0.987012987012987,
+      "families": {
+        "desc": [
+          13,
+          15
+        ],
+        "negation": [
+          10,
+          10
+        ],
+        "logic": [
+          13,
+          17
+        ],
+        "score": [
+          11,
+          12
+        ],
+        "json": [
+          6,
+          6
+        ],
+        "taxonomy": [
+          14,
+          14
+        ],
+        "plausible": [
+          4,
+          4
+        ],
+        "twist": [
+          3,
+          3
+        ],
+        "time": [
+          4,
+          5
+        ],
+        "intent": [
+          2,
+          3
+        ],
+        "compare": [
+          3,
+          4
+        ],
+        "criteria": [
+          4,
+          4
+        ]
+      }
+    },
+    "sst2": {
+      "n": 872,
+      "acc": 0.9575688073394495,
+      "nll": 0.1396011574813463,
+      "brier": 0.07172239838956931,
+      "ece": 0.014292597087151384,
+      "mean_conf": 0.9680821417121712,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.9575688073394495,
+      "cov@0.7": 0.9793577981651376,
+      "acc@0.7": 0.9637002341920374,
+      "cov@0.9": 0.9438073394495413,
+      "acc@0.9": 0.9732685297691372
+    },
+    "mnli": {
+      "n": 1000,
+      "acc": 0.859,
+      "nll": 0.3888388354725572,
+      "brier": 0.21223603757591059,
+      "ece": 0.04788112017512322,
+      "mean_conf": 0.8856889481842518,
+      "cov@0.5": 0.987,
+      "acc@0.5": 0.8662613981762918,
+      "cov@0.7": 0.909,
+      "acc@0.7": 0.8954895489548955,
+      "cov@0.9": 0.655,
+      "acc@0.9": 0.9587786259541985
+    },
+    "cola": {
+      "n": 1000,
+      "acc": 0.759,
+      "nll": 0.5066963657737495,
+      "brier": 0.33443985155856215,
+      "ece": 0.0523103475570679,
+      "mean_conf": 0.791339822769165,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.759,
+      "cov@0.7": 0.764,
+      "acc@0.7": 0.8167539267015707,
+      "cov@0.9": 0.207,
+      "acc@0.9": 0.9516908212560387
+    }
+  },
+  "summary": {
+    "mean_acc": 0.99,
+    "mean_ece": 0.031613957881927515
+  },
+  "history": []
+}
\ No newline at end of file
diff --git a/v0007-support/adapter/README.md b/v0007-support/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007-support/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007-support/adapter/adapter_config.json b/v0007-support/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..4c4d24673a3201f69130c64d59be85707e46cf94
--- /dev/null
+++ b/v0007-support/adapter/adapter_config.json
@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "o_proj",
+    "k_proj",
+    "v_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007-support/adapter/adapter_model.safetensors b/v0007-support/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..69f03c049f8ca0736c2b01ea063a9780d904cf23
--- /dev/null
+++ b/v0007-support/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7946a71e55508a0d6c003df763e4dc48a5a68936214e12c732155b2947c7196d
+size 8745704
diff --git a/v0007-support/config.json b/v0007-support/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5a06253609d2e12a0d57418deca14a0d40ad552
--- /dev/null
+++ b/v0007-support/config.json
@@ -0,0 +1,11 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "parents": [
+    "v0007"
+  ],
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0
+}
\ No newline at end of file
diff --git a/v0007-support/manifest.json b/v0007-support/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..b5d9998c2ed5efd2a8c18eb6449be783ced9a359
--- /dev/null
+++ b/v0007-support/manifest.json
@@ -0,0 +1,57 @@
+{
+  "version": "v0007-support",
+  "parent": "v0007",
+  "parents": [
+    "v0007"
+  ],
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-18T12:52:37+00:00",
+  "tasks": [
+    "banking77",
+    "clinc150",
+    "massive_intent"
+  ],
+  "trained_on": [
+    "banking77",
+    "clinc150",
+    "massive_intent"
+  ],
+  "holdout": [
+    "trec"
+  ],
+  "steps": 1067,
+  "train_examples": 12000,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 8,
+    "loss": "mix",
+    "max_per_task": 4000,
+    "epochs": 1,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {},
+  "unseen_test_uncalibrated": {
+    "trec": {
+      "n": 500,
+      "acc": 0.758,
+      "nll": 0.767611707548029,
+      "brier": 0.37113630112335455,
+      "ece": 0.08773053616285323,
+      "mean_conf": 0.8239709965586662,
+      "cov@0.5": 0.924,
+      "acc@0.5": 0.7835497835497836,
+      "cov@0.7": 0.762,
+      "acc@0.7": 0.8293963254593176,
+      "cov@0.9": 0.474,
+      "acc@0.9": 0.8734177215189873
+    }
+  },
+  "summary": {
+    "mean_acc": null,
+    "mean_ece": null
+  },
+  "history": []
+}
\ No newline at end of file
diff --git a/v0007-typed/adapter/README.md b/v0007-typed/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007-typed/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007-typed/adapter/adapter_config.json b/v0007-typed/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..2db185bf7f3cebb1ea8e4fcd148c15a304d78800
--- /dev/null
+++ b/v0007-typed/adapter/adapter_config.json
@@ -0,0 +1,51 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "k_proj",
+    "down_proj",
+    "up_proj",
+    "o_proj",
+    "gate_proj",
+    "v_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007-typed/adapter/adapter_model.safetensors b/v0007-typed/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..4432d91df12a160e44ce16b721be016dc96c09cf
--- /dev/null
+++ b/v0007-typed/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:149d0ab7101690ea40791ef8eec305fda05ffb73a72fff13260c2ef0827deb88
+size 73911112
diff --git a/v0007-typed/config.json b/v0007-typed/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5a06253609d2e12a0d57418deca14a0d40ad552
--- /dev/null
+++ b/v0007-typed/config.json
@@ -0,0 +1,11 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "parents": [
+    "v0007"
+  ],
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0
+}
\ No newline at end of file
diff --git a/v0007-typed/manifest.json b/v0007-typed/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..bbec5360763910b4a23e8136154b15c4f576d474
--- /dev/null
+++ b/v0007-typed/manifest.json
@@ -0,0 +1,52 @@
+{
+  "version": "v0007-typed",
+  "parent": "v0007",
+  "parents": [
+    "v0007"
+  ],
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-22T11:39:50+00:00",
+  "tasks": [
+    "typed_decisions"
+  ],
+  "trained_on": [
+    "typed_decisions"
+  ],
+  "holdout": [],
+  "steps": 1352,
+  "train_examples": 5400,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 16,
+    "loss": "mix",
+    "max_per_task": 6000,
+    "epochs": 2,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {
+    "typed_decisions": {
+      "n": 300,
+      "acc": 0.85,
+      "nll": 0.7897254475627804,
+      "brier": 0.04430931889174507,
+      "ece": 0.21966197381416955,
+      "mean_conf": 0.6374534544348717,
+      "score_mae": 0.34401526501434937,
+      "cov@0.5": 0.76,
+      "acc@0.5": 0.9342105263157895,
+      "cov@0.7": 0.30666666666666664,
+      "acc@0.7": 0.9891304347826086,
+      "cov@0.9": 0.13666666666666666,
+      "acc@0.9": 1.0
+    }
+  },
+  "unseen_test_uncalibrated": {},
+  "summary": {
+    "mean_acc": 0.85,
+    "mean_ece": 0.21966197381416955
+  },
+  "history": []
+}
\ No newline at end of file
diff --git a/v0007/adapter/README.md b/v0007/adapter/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ece460e8ecf1714cee5aa369eb7a2753c4e14b60
--- /dev/null
+++ b/v0007/adapter/README.md
@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-1.5B-Instruct
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0
\ No newline at end of file
diff --git a/v0007/adapter/adapter_config.json b/v0007/adapter/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..a5f3cc08649cc2ef21b0eae6d5e04c5af99ff15f
--- /dev/null
+++ b/v0007/adapter/adapter_config.json
@@ -0,0 +1,51 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "up_proj",
+    "v_proj",
+    "down_proj",
+    "q_proj",
+    "k_proj",
+    "o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}
\ No newline at end of file
diff --git a/v0007/adapter/adapter_model.safetensors b/v0007/adapter/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..00f233bbab6b3648549015553af07dff249cfdf0
--- /dev/null
+++ b/v0007/adapter/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1440f82c9cf591ec770b21d1f2f30b542ea8a368ea875b11a74027bba549d199
+size 73911112
diff --git a/v0007/config.json b/v0007/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..fbf9c143b39803e63703f5610a50d882861e0fbb
--- /dev/null
+++ b/v0007/config.json
@@ -0,0 +1,13 @@
+{
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "adapter": "adapter",
+  "max_state_tokens": 700,
+  "max_len": 1024,
+  "temperature": 1.0194810628890991,
+  "temperature_by_k": {
+    "2-2": 0.9909,
+    "3-5": 1.0355,
+    "6-20": 1.0193
+  }
+}
\ No newline at end of file
diff --git a/v0007/eval_test.json b/v0007/eval_test.json
new file mode 100644
index 0000000000000000000000000000000000000000..e9108de8dad97e40ca60e770db5feac764457181
--- /dev/null
+++ b/v0007/eval_test.json
@@ -0,0 +1,1017 @@
+{
+  "ag_news": {
+    "n": 500,
+    "acc": 0.904,
+    "nll": 0.2952045052830233,
+    "brier": 0.14562407385514445,
+    "ece": 0.03137661844491955,
+    "mean_conf": 0.9168671212792396,
+    "cov@0.5": 0.992,
+    "acc@0.5": 0.9092741935483871,
+    "cov@0.7": 0.91,
+    "acc@0.7": 0.945054945054945,
+    "cov@0.9": 0.762,
+    "acc@0.9": 0.963254593175853
+  },
+  "boolq": {
+    "n": 500,
+    "acc": 0.822,
+    "nll": 0.38821684995019157,
+    "brier": 0.24384808029659086,
+    "ece": 0.044399895310401956,
+    "mean_conf": 0.8563220142126083,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.822,
+    "cov@0.7": 0.866,
+    "acc@0.7": 0.8729792147806005,
+    "cov@0.9": 0.486,
+    "acc@0.9": 0.9465020576131687
+  },
+  "commonsense_qa": {
+    "n": 500,
+    "acc": 0.792,
+    "nll": 0.6138367100376695,
+    "brier": 0.31621131048467216,
+    "ece": 0.04827169889211652,
+    "mean_conf": 0.795568238556385,
+    "cov@0.5": 0.896,
+    "acc@0.5": 0.8125,
+    "cov@0.7": 0.704,
+    "acc@0.7": 0.8835227272727273,
+    "cov@0.9": 0.422,
+    "acc@0.9": 0.933649289099526
+  },
+  "copa": {
+    "n": 100,
+    "acc": 0.93,
+    "nll": 0.16677208632899737,
+    "brier": 0.09258839384263035,
+    "ece": 0.04937064290046691,
+    "mean_conf": 0.9120474457740784,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.93,
+    "cov@0.7": 0.94,
+    "acc@0.7": 0.9680851063829787,
+    "cov@0.9": 0.71,
+    "acc@0.9": 1.0
+  },
+  "dbpedia": {
+    "n": 500,
+    "acc": 0.978,
+    "nll": 0.08366780531041144,
+    "brier": 0.034640499350234014,
+    "ece": 0.01524236822128298,
+    "mean_conf": 0.9774349620342254,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.978,
+    "cov@0.7": 0.986,
+    "acc@0.7": 0.9858012170385395,
+    "cov@0.9": 0.96,
+    "acc@0.9": 0.99375
+  },
+  "emotion": {
+    "n": 500,
+    "acc": 0.794,
+    "nll": 0.5465345826535903,
+    "brier": 0.28449990068829645,
+    "ece": 0.048147320568561594,
+    "mean_conf": 0.8112419300675392,
+    "cov@0.5": 0.936,
+    "acc@0.5": 0.8247863247863247,
+    "cov@0.7": 0.744,
+    "acc@0.7": 0.8978494623655914,
+    "cov@0.9": 0.426,
+    "acc@0.9": 0.9765258215962441
+  },
+  "facts": {
+    "n": 30,
+    "acc": 0.9,
+    "nll": 0.2075752172314097,
+    "brier": 0.12133284522807337,
+    "ece": 0.09556647539138795,
+    "mean_conf": 0.9290410478909811,
+    "cov@0.5": 0.9666666666666667,
+    "acc@0.5": 0.9310344827586207,
+    "cov@0.7": 0.9,
+    "acc@0.7": 0.9629629629629629,
+    "cov@0.9": 0.8666666666666667,
+    "acc@0.9": 0.9615384615384616
+  },
+  "fits": {
+    "n": 207,
+    "acc": 0.8985507246376812,
+    "nll": 0.4405018217697743,
+    "brier": 0.07160491583881325,
+    "ece": 0.08918811791185018,
+    "mean_conf": 0.8246165606134755,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.8985507246376812,
+    "cov@0.7": 0.8405797101449275,
+    "acc@0.7": 0.9425287356321839,
+    "cov@0.9": 0.3333333333333333,
+    "acc@0.9": 0.9855072463768116
+  },
+  "imdb": {
+    "n": 500,
+    "acc": 0.964,
+    "nll": 0.10838915214226351,
+    "brier": 0.05855760938480377,
+    "ece": 0.017840196132660004,
+    "mean_conf": 0.9783704316616059,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.964,
+    "cov@0.7": 0.988,
+    "acc@0.7": 0.9696356275303644,
+    "cov@0.9": 0.944,
+    "acc@0.9": 0.9809322033898306
+  },
+  "jailbreak": {
+    "n": 262,
+    "acc": 0.9847328244274809,
+    "nll": 0.053944837311566685,
+    "brier": 0.02485429640642806,
+    "ece": 0.027110507470050817,
+    "mean_conf": 0.9702985978308525,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.9847328244274809,
+    "cov@0.7": 0.9885496183206107,
+    "acc@0.7": 0.9884169884169884,
+    "cov@0.9": 0.9541984732824428,
+    "acc@0.9": 0.996
+  },
+  "mnli": {
+    "n": 500,
+    "acc": 0.872,
+    "nll": 0.37104603358653826,
+    "brier": 0.2035345314241094,
+    "ece": 0.04369585227966309,
+    "mean_conf": 0.8777530242204666,
+    "cov@0.5": 0.976,
+    "acc@0.5": 0.8770491803278688,
+    "cov@0.7": 0.888,
+    "acc@0.7": 0.9054054054054054,
+    "cov@0.9": 0.634,
+    "acc@0.9": 0.9589905362776026
+  },
+  "mrpc": {
+    "n": 408,
+    "acc": 0.8382352941176471,
+    "nll": 0.3661221842177323,
+    "brier": 0.2271248635333212,
+    "ece": 0.0363529923499799,
+    "mean_conf": 0.8275741233545191,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.8382352941176471,
+    "cov@0.7": 0.8112745098039216,
+    "acc@0.7": 0.9003021148036254,
+    "cov@0.9": 0.40931372549019607,
+    "acc@0.9": 0.9640718562874252
+  },
+  "openbookqa": {
+    "n": 500,
+    "acc": 0.844,
+    "nll": 0.4436366222899719,
+    "brier": 0.23142732591676246,
+    "ece": 0.03274157041311261,
+    "mean_conf": 0.8326144033074379,
+    "cov@0.5": 0.92,
+    "acc@0.5": 0.8717391304347826,
+    "cov@0.7": 0.778,
+    "acc@0.7": 0.9203084832904884,
+    "cov@0.9": 0.544,
+    "acc@0.9": 0.9742647058823529
+  },
+  "paws": {
+    "n": 500,
+    "acc": 0.916,
+    "nll": 0.20104104238901066,
+    "brier": 0.12062216912045075,
+    "ece": 0.03848682808876038,
+    "mean_conf": 0.9170866451263427,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.916,
+    "cov@0.7": 0.942,
+    "acc@0.7": 0.9384288747346072,
+    "cov@0.9": 0.736,
+    "acc@0.9": 0.9918478260869565
+  },
+  "qnli": {
+    "n": 500,
+    "acc": 0.906,
+    "nll": 0.26933529241738075,
+    "brier": 0.15768804149138618,
+    "ece": 0.02181814336776728,
+    "mean_conf": 0.8996418695449829,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.906,
+    "cov@0.7": 0.918,
+    "acc@0.7": 0.9215686274509803,
+    "cov@0.9": 0.688,
+    "acc@0.9": 0.9563953488372093
+  },
+  "read": {
+    "n": 400,
+    "acc": 1.0,
+    "nll": 0.013971105909523766,
+    "brier": 0.0006018934900539811,
+    "ece": 0.01377617731690402,
+    "mean_conf": 0.9862238226830959,
+    "cov@0.5": 1.0,
+    "acc@0.5": 1.0,
+    "cov@0.7": 1.0,
+    "acc@0.7": 1.0,
+    "cov@0.9": 0.9975,
+    "acc@0.9": 1.0
+  },
+  "rte": {
+    "n": 277,
+    "acc": 0.855595667870036,
+    "nll": 0.3405857359516095,
+    "brier": 0.20732145191787296,
+    "ece": 0.04871643859126507,
+    "mean_conf": 0.8968358597170145,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.855595667870036,
+    "cov@0.7": 0.9169675090252708,
+    "acc@0.7": 0.889763779527559,
+    "cov@0.9": 0.6895306859205776,
+    "acc@0.9": 0.93717277486911
+  },
+  "sciq": {
+    "n": 500,
+    "acc": 0.974,
+    "nll": 0.06468619566328364,
+    "brier": 0.03523030876753992,
+    "ece": 0.020325540423393212,
+    "mean_conf": 0.9750150481462478,
+    "cov@0.5": 0.996,
+    "acc@0.5": 0.9759036144578314,
+    "cov@0.7": 0.972,
+    "acc@0.7": 0.9876543209876543,
+    "cov@0.9": 0.94,
+    "acc@0.9": 0.997872340425532
+  },
+  "sst2": {
+    "n": 500,
+    "acc": 0.952,
+    "nll": 0.15055633344092564,
+    "brier": 0.0785655014897793,
+    "ece": 0.015440203905105584,
+    "mean_conf": 0.9604808611869812,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.952,
+    "cov@0.7": 0.978,
+    "acc@0.7": 0.9611451942740287,
+    "cov@0.9": 0.894,
+    "acc@0.9": 0.9731543624161074
+  },
+  "swag": {
+    "n": 500,
+    "acc": 0.836,
+    "nll": 0.46389228138841154,
+    "brier": 0.23638607587807828,
+    "ece": 0.05372774058580401,
+    "mean_conf": 0.792816475212574,
+    "cov@0.5": 0.908,
+    "acc@0.5": 0.8854625550660793,
+    "cov@0.7": 0.722,
+    "acc@0.7": 0.9390581717451524,
+    "cov@0.9": 0.384,
+    "acc@0.9": 0.984375
+  },
+  "tweet_emoji": {
+    "n": 500,
+    "acc": 0.412,
+    "nll": 2.0648460126740087,
+    "brier": 0.7092518610639151,
+    "ece": 0.05932182088494301,
+    "mean_conf": 0.3580406456887722,
+    "cov@0.5": 0.244,
+    "acc@0.5": 0.8934426229508197,
+    "cov@0.7": 0.192,
+    "acc@0.7": 0.9583333333333334,
+    "cov@0.9": 0.072,
+    "acc@0.9": 0.9444444444444444
+  },
+  "tweet_hate": {
+    "n": 500,
+    "acc": 0.46,
+    "nll": 1.2138538786047812,
+    "brier": 0.8031272788007013,
+    "ece": 0.39924867236614225,
+    "mean_conf": 0.8592486723661422,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.46,
+    "cov@0.7": 0.892,
+    "acc@0.7": 0.47533632286995514,
+    "cov@0.9": 0.486,
+    "acc@0.9": 0.5432098765432098
+  },
+  "tweet_irony": {
+    "n": 500,
+    "acc": 0.766,
+    "nll": 0.4896152915870031,
+    "brier": 0.3183358836195773,
+    "ece": 0.06906565117836,
+    "mean_conf": 0.7235815870761871,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.766,
+    "cov@0.7": 0.578,
+    "acc@0.7": 0.8788927335640139,
+    "cov@0.9": 0.092,
+    "acc@0.9": 0.9565217391304348
+  },
+  "tweet_offensive": {
+    "n": 500,
+    "acc": 0.83,
+    "nll": 0.3822173999368801,
+    "brier": 0.2395516581467213,
+    "ece": 0.03853944563865664,
+    "mean_conf": 0.8170348312854767,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.83,
+    "cov@0.7": 0.8,
+    "acc@0.7": 0.9,
+    "cov@0.9": 0.346,
+    "acc@0.9": 0.976878612716763
+  },
+  "tweet_sentiment": {
+    "n": 500,
+    "acc": 0.738,
+    "nll": 0.58249910191355,
+    "brier": 0.3534792677090853,
+    "ece": 0.03204844713211058,
+    "mean_conf": 0.733830077290535,
+    "score_mae": 0.335998848663643,
+    "cov@0.5": 0.946,
+    "acc@0.5": 0.7547568710359408,
+    "cov@0.7": 0.61,
+    "acc@0.7": 0.8459016393442623,
+    "cov@0.9": 0.13,
+    "acc@0.9": 0.9538461538461539
+  },
+  "yahoo": {
+    "n": 500,
+    "acc": 0.746,
+    "nll": 0.835068198749748,
+    "brier": 0.36347902333918586,
+    "ece": 0.05916145133972168,
+    "mean_conf": 0.7927027177810669,
+    "cov@0.5": 0.864,
+    "acc@0.5": 0.8125,
+    "cov@0.7": 0.718,
+    "acc@0.7": 0.8523676880222841,
+    "cov@0.9": 0.444,
+    "acc@0.9": 0.9504504504504504
+  },
+  "yelp": {
+    "n": 500,
+    "acc": 0.672,
+    "nll": 0.7321823700995426,
+    "brier": 0.4442138012250366,
+    "ece": 0.07660222238302232,
+    "mean_conf": 0.7385406532883644,
+    "score_mae": 0.39326068315636076,
+    "cov@0.5": 0.952,
+    "acc@0.5": 0.6848739495798319,
+    "cov@0.7": 0.634,
+    "acc@0.7": 0.7444794952681388,
+    "cov@0.9": 0.148,
+    "acc@0.9": 0.9324324324324325
+  },
+  "anli": {
+    "n": 500,
+    "acc": 0.532,
+    "nll": 1.1065353592560714,
+    "brier": 0.663176574205106,
+    "ece": 0.22013714534044268,
+    "mean_conf": 0.7495694995522499,
+    "cov@0.5": 0.904,
+    "acc@0.5": 0.5398230088495575,
+    "cov@0.7": 0.652,
+    "acc@0.7": 0.5858895705521472,
+    "cov@0.9": 0.222,
+    "acc@0.9": 0.6036036036036037
+  },
+  "winogrande": {
+    "n": 500,
+    "acc": 0.676,
+    "nll": 0.6965373426873016,
+    "brier": 0.4634204682399545,
+    "ece": 0.148663760304451,
+    "mean_conf": 0.8135936650037765,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.676,
+    "cov@0.7": 0.766,
+    "acc@0.7": 0.7154046997389034,
+    "cov@0.9": 0.384,
+    "acc@0.9": 0.7916666666666666
+  },
+  "hellaswag": {
+    "n": 500,
+    "acc": 0.854,
+    "nll": 0.40399329923562644,
+    "brier": 0.21240321793802722,
+    "ece": 0.0430728812813759,
+    "mean_conf": 0.8583100009560585,
+    "cov@0.5": 0.954,
+    "acc@0.5": 0.870020964360587,
+    "cov@0.7": 0.828,
+    "acc@0.7": 0.9202898550724637,
+    "cov@0.9": 0.578,
+    "acc@0.9": 0.9826989619377162
+  },
+  "race": {
+    "n": 500,
+    "acc": 0.79,
+    "nll": 0.617019975755489,
+    "brier": 0.3096390749225188,
+    "ece": 0.05949199843406676,
+    "mean_conf": 0.8478295928239823,
+    "cov@0.5": 0.934,
+    "acc@0.5": 0.8222698072805139,
+    "cov@0.7": 0.796,
+    "acc@0.7": 0.864321608040201,
+    "cov@0.9": 0.574,
+    "acc@0.9": 0.926829268292683
+  },
+  "scitail": {
+    "n": 500,
+    "acc": 0.93,
+    "nll": 0.1743719026375374,
+    "brier": 0.101472771827948,
+    "ece": 0.02574878227710722,
+    "mean_conf": 0.9325147467851639,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.93,
+    "cov@0.7": 0.942,
+    "acc@0.7": 0.9554140127388535,
+    "cov@0.9": 0.808,
+    "acc@0.9": 0.9801980198019802
+  },
+  "qqp": {
+    "n": 500,
+    "acc": 0.848,
+    "nll": 0.3165067298532316,
+    "brier": 0.20145011381454622,
+    "ece": 0.04761451995372773,
+    "mean_conf": 0.8586561111211777,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.848,
+    "cov@0.7": 0.83,
+    "acc@0.7": 0.9156626506024096,
+    "cov@0.9": 0.526,
+    "acc@0.9": 0.9771863117870723
+  },
+  "stsb": {
+    "n": 500,
+    "acc": 0.61,
+    "nll": 1.0147768329027056,
+    "brier": 0.2755420434300338,
+    "ece": 0.06956788003444671,
+    "mean_conf": 0.5404321199655533,
+    "score_mae": 0.5215406965478323,
+    "cov@0.5": 0.59,
+    "acc@0.5": 0.6474576271186441,
+    "cov@0.7": 0.09,
+    "acc@0.7": 0.9111111111111111,
+    "cov@0.9": 0.002,
+    "acc@0.9": 1.0
+  },
+  "toxic": {
+    "n": 500,
+    "acc": 0.864,
+    "nll": 0.3133165936564978,
+    "brier": 0.19149039831133785,
+    "ece": 0.031214740157127406,
+    "mean_conf": 0.8684242066144944,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.864,
+    "cov@0.7": 0.87,
+    "acc@0.7": 0.9126436781609195,
+    "cov@0.9": 0.586,
+    "acc@0.9": 0.9692832764505119
+  },
+  "stance_abortion": {
+    "n": 280,
+    "acc": 0.6392857142857142,
+    "nll": 0.7371370510055495,
+    "brier": 0.4538894518081303,
+    "ece": 0.11449087389877866,
+    "mean_conf": 0.7491982987948826,
+    "cov@0.5": 0.9285714285714286,
+    "acc@0.5": 0.6615384615384615,
+    "cov@0.7": 0.6607142857142857,
+    "acc@0.7": 0.7567567567567568,
+    "cov@0.9": 0.14285714285714285,
+    "acc@0.9": 0.95
+  },
+  "stance_atheism": {
+    "n": 220,
+    "acc": 0.759090909090909,
+    "nll": 0.5504364743881415,
+    "brier": 0.3295810547096517,
+    "ece": 0.06360254734754561,
+    "mean_conf": 0.7973180409182202,
+    "cov@0.5": 0.95,
+    "acc@0.5": 0.7751196172248804,
+    "cov@0.7": 0.7181818181818181,
+    "acc@0.7": 0.8481012658227848,
+    "cov@0.9": 0.33181818181818185,
+    "acc@0.9": 0.9178082191780822
+  },
+  "stance_feminist": {
+    "n": 285,
+    "acc": 0.7087719298245614,
+    "nll": 0.6968129609369124,
+    "brier": 0.41241720905885676,
+    "ece": 0.07303431546478939,
+    "mean_conf": 0.7419619864539096,
+    "cov@0.5": 0.9122807017543859,
+    "acc@0.5": 0.7192307692307692,
+    "cov@0.7": 0.6350877192982456,
+    "acc@0.7": 0.8011049723756906,
+    "cov@0.9": 0.15789473684210525,
+    "acc@0.9": 0.9555555555555556
+  },
+  "stance_hillary": {
+    "n": 295,
+    "acc": 0.7457627118644068,
+    "nll": 0.5526740804818927,
+    "brier": 0.32719331452723405,
+    "ece": 0.06370943774611264,
+    "mean_conf": 0.7728302831366911,
+    "cov@0.5": 0.9491525423728814,
+    "acc@0.5": 0.7714285714285715,
+    "cov@0.7": 0.6813559322033899,
+    "acc@0.7": 0.8756218905472637,
+    "cov@0.9": 0.27796610169491526,
+    "acc@0.9": 0.9634146341463414
+  },
+  "match": {
+    "n": 500,
+    "acc": 0.99,
+    "nll": 0.031114264246498352,
+    "brier": 0.012187918203564489,
+    "ece": 0.015517388939857501,
+    "mean_conf": 0.9813971043825149,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.99,
+    "cov@0.7": 0.992,
+    "acc@0.7": 0.9959677419354839,
+    "cov@0.9": 0.968,
+    "acc@0.9": 0.9979338842975206
+  },
+  "reason": {
+    "n": 500,
+    "acc": 0.946,
+    "nll": 0.16985629533349725,
+    "brier": 0.08885515226028988,
+    "ece": 0.05641485232114797,
+    "mean_conf": 0.8917767191529274,
+    "cov@0.5": 0.974,
+    "acc@0.5": 0.9630390143737166,
+    "cov@0.7": 0.852,
+    "acc@0.7": 0.9835680751173709,
+    "cov@0.9": 0.736,
+    "acc@0.9": 0.9945652173913043
+  },
+  "formality": {
+    "n": 500,
+    "acc": 0.634,
+    "nll": 1.0050057317205758,
+    "brier": 0.2292107567592736,
+    "ece": 0.11595214062929157,
+    "mean_conf": 0.5217105874419212,
+    "score_mae": 0.4580765849482268,
+    "cov@0.5": 0.622,
+    "acc@0.5": 0.6816720257234726,
+    "cov@0.7": 0.006,
+    "acc@0.7": 0.3333333333333333,
+    "cov@0.9": 0.0,
+    "acc@0.9": NaN
+  },
+  "politeness": {
+    "n": 500,
+    "acc": 0.884,
+    "nll": 0.30688122991376177,
+    "brier": 0.16651920171747073,
+    "ece": 0.04738429147005082,
+    "mean_conf": 0.8626298355460167,
+    "score_mae": 0.18424055973393844,
+    "cov@0.5": 0.964,
+    "acc@0.5": 0.9004149377593361,
+    "cov@0.7": 0.834,
+    "acc@0.7": 0.947242206235012,
+    "cov@0.9": 0.574,
+    "acc@0.9": 0.9895470383275261
+  },
+  "strategyqa": {
+    "n": 500,
+    "acc": 0.656,
+    "nll": 0.5973554241686542,
+    "brier": 0.41449589929170566,
+    "ece": 0.04992203032970431,
+    "mean_conf": 0.6695813618898392,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.656,
+    "cov@0.7": 0.346,
+    "acc@0.7": 0.8034682080924855,
+    "cov@0.9": 0.06,
+    "acc@0.9": 0.9666666666666667
+  },
+  "vitaminc": {
+    "n": 500,
+    "acc": 0.816,
+    "nll": 0.5219780039570592,
+    "brier": 0.28570030051715434,
+    "ece": 0.039564442515373215,
+    "mean_conf": 0.823793786406517,
+    "cov@0.5": 0.962,
+    "acc@0.5": 0.8295218295218295,
+    "cov@0.7": 0.802,
+    "acc@0.7": 0.8678304239401496,
+    "cov@0.9": 0.404,
+    "acc@0.9": 0.9504950495049505
+  },
+  "ruletaker": {
+    "n": 500,
+    "acc": 0.82,
+    "nll": 0.38612913783564756,
+    "brier": 0.252674128497622,
+    "ece": 0.05111093926429748,
+    "mean_conf": 0.7891134247779846,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.82,
+    "cov@0.7": 0.658,
+    "acc@0.7": 0.9148936170212766,
+    "cov@0.9": 0.39,
+    "acc@0.9": 0.9846153846153847
+  },
+  "proofwriter": {
+    "n": 500,
+    "acc": 0.846,
+    "nll": 0.42337107551622927,
+    "brier": 0.2461669688565621,
+    "ece": 0.0678113833665848,
+    "mean_conf": 0.8122986115217209,
+    "cov@0.5": 0.972,
+    "acc@0.5": 0.8415637860082305,
+    "cov@0.7": 0.772,
+    "acc@0.7": 0.8860103626943006,
+    "cov@0.9": 0.394,
+    "acc@0.9": 0.9746192893401016
+  },
+  "folio": {
+    "n": 203,
+    "acc": 0.6600985221674877,
+    "nll": 0.7939969248617568,
+    "brier": 0.46891455062020077,
+    "ece": 0.08365710674248306,
+    "mean_conf": 0.6265559084896971,
+    "cov@0.5": 0.7339901477832512,
+    "acc@0.5": 0.7181208053691275,
+    "cov@0.7": 0.3054187192118227,
+    "acc@0.7": 0.8709677419354839,
+    "cov@0.9": 0.04433497536945813,
+    "acc@0.9": 0.8888888888888888
+  },
+  "logiqa": {
+    "n": 500,
+    "acc": 0.624,
+    "nll": 0.6597909275993277,
+    "brier": 0.46379411691209055,
+    "ece": 0.032960258245468124,
+    "mean_conf": 0.6369408411979676,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.624,
+    "cov@0.7": 0.27,
+    "acc@0.7": 0.7333333333333333,
+    "cov@0.9": 0.01,
+    "acc@0.9": 0.8
+  },
+  "tracie": {
+    "n": 500,
+    "acc": 0.712,
+    "nll": 0.5713794439647892,
+    "brier": 0.3878056407184628,
+    "ece": 0.054143391370773314,
+    "mean_conf": 0.7030656831264496,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.712,
+    "cov@0.7": 0.538,
+    "acc@0.7": 0.8066914498141264,
+    "cov@0.9": 0.032,
+    "acc@0.9": 0.9375
+  },
+  "temporal_nli": {
+    "n": 500,
+    "acc": 0.832,
+    "nll": 0.4014951934609327,
+    "brier": 0.23685619378316686,
+    "ece": 0.04163959693908693,
+    "mean_conf": 0.7919499685764313,
+    "cov@0.5": 0.988,
+    "acc@0.5": 0.8380566801619433,
+    "cov@0.7": 0.754,
+    "acc@0.7": 0.9018567639257294,
+    "cov@0.9": 0.232,
+    "acc@0.9": 1.0
+  },
+  "piqa": {
+    "n": 500,
+    "acc": 0.826,
+    "nll": 0.3592462708765769,
+    "brier": 0.22998577175129606,
+    "ece": 0.027211105823516855,
+    "mean_conf": 0.8398146278858185,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.826,
+    "cov@0.7": 0.788,
+    "acc@0.7": 0.9010152284263959,
+    "cov@0.9": 0.462,
+    "acc@0.9": 0.9567099567099567
+  },
+  "siqa": {
+    "n": 500,
+    "acc": 0.744,
+    "nll": 0.5927271855927616,
+    "brier": 0.34397084336114087,
+    "ece": 0.056237839221954314,
+    "mean_conf": 0.8002378392219544,
+    "cov@0.5": 0.948,
+    "acc@0.5": 0.770042194092827,
+    "cov@0.7": 0.738,
+    "acc@0.7": 0.8536585365853658,
+    "cov@0.9": 0.386,
+    "acc@0.9": 0.9378238341968912
+  },
+  "clutrr": {
+    "n": 500,
+    "acc": 0.342,
+    "nll": 1.7583232741068568,
+    "brier": 0.784316384590205,
+    "ece": 0.18587388846278188,
+    "mean_conf": 0.527041629999876,
+    "cov@0.5": 0.55,
+    "acc@0.5": 0.41818181818181815,
+    "cov@0.7": 0.106,
+    "acc@0.7": 0.6037735849056604,
+    "cov@0.9": 0.008,
+    "acc@0.9": 1.0
+  },
+  "gsm8k": {
+    "n": 500,
+    "acc": 0.69,
+    "nll": 0.7243987473415922,
+    "brier": 0.3908694153304782,
+    "ece": 0.05401642280817035,
+    "mean_conf": 0.6785197833180427,
+    "cov@0.5": 0.746,
+    "acc@0.5": 0.7989276139410187,
+    "cov@0.7": 0.48,
+    "acc@0.7": 0.9208333333333333,
+    "cov@0.9": 0.19,
+    "acc@0.9": 1.0
+  },
+  "svamp": {
+    "n": 300,
+    "acc": 0.6333333333333333,
+    "nll": 0.8223909804125111,
+    "brier": 0.46969116350636153,
+    "ece": 0.05721060862143834,
+    "mean_conf": 0.6183922579884529,
+    "cov@0.5": 0.6966666666666667,
+    "acc@0.5": 0.7081339712918661,
+    "cov@0.7": 0.3,
+    "acc@0.7": 0.8444444444444444,
+    "cov@0.9": 0.08333333333333333,
+    "acc@0.9": 0.96
+  },
+  "aqua": {
+    "n": 247,
+    "acc": 0.3441295546558704,
+    "nll": 1.4699616183019182,
+    "brier": 0.7386751804022209,
+    "ece": 0.047319990299973885,
+    "mean_conf": 0.36171867695414583,
+    "cov@0.5": 0.08502024291497975,
+    "acc@0.5": 0.7142857142857143,
+    "cov@0.7": 0.03643724696356275,
+    "acc@0.7": 0.7777777777777778,
+    "cov@0.9": 0.008097165991902834,
+    "acc@0.9": 1.0
+  },
+  "probe": {
+    "n": 97,
+    "acc": 0.8969072164948454,
+    "nll": 0.20803647221069066,
+    "brier": 0.1236771392560089,
+    "ece": 0.07140122492288804,
+    "mean_conf": 0.9083748072693029,
+    "score_mae": 0.1922900263549915,
+    "cov@0.5": 0.9896907216494846,
+    "acc@0.5": 0.90625,
+    "cov@0.7": 0.9072164948453608,
+    "acc@0.7": 0.9545454545454546,
+    "cov@0.9": 0.7525773195876289,
+    "acc@0.9": 0.9863013698630136
+  },
+  "bbh": {
+    "n": 500,
+    "acc": 0.5,
+    "nll": 1.126052411237626,
+    "brier": 0.6115770776560393,
+    "ece": 0.09403421479463578,
+    "mean_conf": 0.5716841769814491,
+    "cov@0.5": 0.62,
+    "acc@0.5": 0.6129032258064516,
+    "cov@0.7": 0.268,
+    "acc@0.7": 0.6865671641791045,
+    "cov@0.9": 0.07,
+    "acc@0.9": 0.7428571428571429
+  },
+  "cola": {
+    "n": 500,
+    "acc": 0.75,
+    "nll": 0.5091424880432279,
+    "brier": 0.33440911474778157,
+    "ece": 0.055637175679206806,
+    "mean_conf": 0.7200385755300522,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.75,
+    "cov@0.7": 0.6,
+    "acc@0.7": 0.8333333333333334,
+    "cov@0.9": 0.012,
+    "acc@0.9": 1.0
+  },
+  "wic": {
+    "n": 500,
+    "acc": 0.59,
+    "nll": 0.6902373164882011,
+    "brier": 0.4917619322792045,
+    "ece": 0.10841803991794587,
+    "mean_conf": 0.6887881902456283,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.59,
+    "cov@0.7": 0.454,
+    "acc@0.7": 0.6696035242290749,
+    "cov@0.9": 0.022,
+    "acc@0.9": 0.8181818181818182
+  },
+  "subj": {
+    "n": 500,
+    "acc": 0.684,
+    "nll": 0.5901592380772634,
+    "brier": 0.4103341799810123,
+    "ece": 0.12139748227596285,
+    "mean_conf": 0.7990830732584,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.684,
+    "cov@0.7": 0.738,
+    "acc@0.7": 0.7452574525745257,
+    "cov@0.9": 0.3,
+    "acc@0.9": 0.94
+  },
+  "spam": {
+    "n": 500,
+    "acc": 0.764,
+    "nll": 0.4650097482775654,
+    "brier": 0.3136384680751385,
+    "ece": 0.0984970170259476,
+    "mean_conf": 0.8226811863183975,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.764,
+    "cov@0.7": 0.82,
+    "acc@0.7": 0.8121951219512196,
+    "cov@0.9": 0.374,
+    "acc@0.9": 0.9679144385026738
+  },
+  "counterfactual": {
+    "n": 500,
+    "acc": 0.82,
+    "nll": 0.4700970053735378,
+    "brier": 0.2962126097422747,
+    "ece": 0.09646137535572051,
+    "mean_conf": 0.7235386246442794,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.82,
+    "cov@0.7": 0.636,
+    "acc@0.7": 0.8867924528301887,
+    "cov@0.9": 0.008,
+    "acc@0.9": 0.75
+  },
+  "cb": {
+    "n": 56,
+    "acc": 0.875,
+    "nll": 0.38689621339204633,
+    "brier": 0.1985667895578788,
+    "ece": 0.12361876879419596,
+    "mean_conf": 0.7997828998735973,
+    "cov@0.5": 0.9821428571428571,
+    "acc@0.5": 0.8909090909090909,
+    "cov@0.7": 0.75,
+    "acc@0.7": 0.9761904761904762,
+    "cov@0.9": 0.2857142857142857,
+    "acc@0.9": 1.0
+  },
+  "arc_challenge": {
+    "n": 500,
+    "acc": 0.786,
+    "nll": 0.594010773816455,
+    "brier": 0.3166836699331607,
+    "ece": 0.04794279056787492,
+    "mean_conf": 0.7890923760533333,
+    "cov@0.5": 0.882,
+    "acc@0.5": 0.8276643990929705,
+    "cov@0.7": 0.69,
+    "acc@0.7": 0.881159420289855,
+    "cov@0.9": 0.418,
+    "acc@0.9": 0.9665071770334929
+  },
+  "stance_climate": {
+    "n": 169,
+    "acc": 0.7100591715976331,
+    "nll": 0.7974690104995089,
+    "brier": 0.43555373727561586,
+    "ece": 0.081085235938518,
+    "mean_conf": 0.7026936018608025,
+    "cov@0.5": 0.8579881656804734,
+    "acc@0.5": 0.7241379310344828,
+    "cov@0.7": 0.5029585798816568,
+    "acc@0.7": 0.8941176470588236,
+    "cov@0.9": 0.11834319526627218,
+    "acc@0.9": 0.95
+  },
+  "trec": {
+    "n": 500,
+    "acc": 0.728,
+    "nll": 0.7760585628620453,
+    "brier": 0.3906730053560261,
+    "ece": 0.044884639918804176,
+    "mean_conf": 0.7680628853440284,
+    "cov@0.5": 0.886,
+    "acc@0.5": 0.781038374717833,
+    "cov@0.7": 0.66,
+    "acc@0.7": 0.8333333333333334,
+    "cov@0.9": 0.348,
+    "acc@0.9": 0.8908045977011494
+  },
+  "sst5": {
+    "n": 500,
+    "acc": 0.588,
+    "nll": 0.962300751103116,
+    "brier": 0.5442527844220892,
+    "ece": 0.07062593251466752,
+    "mean_conf": 0.5504769374728202,
+    "score_mae": 0.533001714671962,
+    "cov@0.5": 0.652,
+    "acc@0.5": 0.6595092024539877,
+    "cov@0.7": 0.112,
+    "acc@0.7": 0.7678571428571429,
+    "cov@0.9": 0.002,
+    "acc@0.9": 1.0
+  },
+  "fin_sentiment": {
+    "n": 500,
+    "acc": 0.806,
+    "nll": 0.48616881277175833,
+    "brier": 0.28441906000643985,
+    "ece": 0.07165524899959566,
+    "mean_conf": 0.7506626476049423,
+    "cov@0.5": 0.986,
+    "acc@0.5": 0.8133874239350912,
+    "cov@0.7": 0.66,
+    "acc@0.7": 0.9,
+    "cov@0.9": 0.108,
+    "acc@0.9": 0.9814814814814815
+  },
+  "arc_easy": {
+    "n": 500,
+    "acc": 0.9,
+    "nll": 0.3002959193224214,
+    "brier": 0.15472624642094474,
+    "ece": 0.018672479689121238,
+    "mean_conf": 0.8853230164647102,
+    "cov@0.5": 0.954,
+    "acc@0.5": 0.9182389937106918,
+    "cov@0.7": 0.852,
+    "acc@0.7": 0.9530516431924883,
+    "cov@0.9": 0.706,
+    "acc@0.9": 0.9745042492917847
+  },
+  "newsgroups": {
+    "n": 500,
+    "acc": 0.648,
+    "nll": 1.2328933369252837,
+    "brier": 0.46483539399426466,
+    "ece": 0.060195389032363905,
+    "mean_conf": 0.6613510022163391,
+    "cov@0.5": 0.69,
+    "acc@0.5": 0.8260869565217391,
+    "cov@0.7": 0.51,
+    "acc@0.7": 0.8862745098039215,
+    "cov@0.9": 0.234,
+    "acc@0.9": 0.9829059829059829
+  }
+}
\ No newline at end of file
diff --git a/v0007/eval_unseen.json b/v0007/eval_unseen.json
new file mode 100644
index 0000000000000000000000000000000000000000..101a44aa0ef497658d5090e34739565c924fb414
--- /dev/null
+++ b/v0007/eval_unseen.json
@@ -0,0 +1,214 @@
+{
+  "probe": {
+    "n": 97,
+    "acc": 0.8969072164948454,
+    "nll": 0.20803647221069066,
+    "brier": 0.1236771392560089,
+    "ece": 0.07140122492288804,
+    "mean_conf": 0.9083748072693029,
+    "score_mae": 0.1922900263549915,
+    "cov@0.5": 0.9896907216494846,
+    "acc@0.5": 0.90625,
+    "cov@0.7": 0.9072164948453608,
+    "acc@0.7": 0.9545454545454546,
+    "cov@0.9": 0.7525773195876289,
+    "acc@0.9": 0.9863013698630136
+  },
+  "bbh": {
+    "n": 1000,
+    "acc": 0.518,
+    "nll": 1.107520868909578,
+    "brier": 0.6032782669998853,
+    "ece": 0.06898981089144945,
+    "mean_conf": 0.5686697762086987,
+    "cov@0.5": 0.628,
+    "acc@0.5": 0.6146496815286624,
+    "cov@0.7": 0.267,
+    "acc@0.7": 0.700374531835206,
+    "cov@0.9": 0.068,
+    "acc@0.9": 0.8235294117647058
+  },
+  "cola": {
+    "n": 1000,
+    "acc": 0.75,
+    "nll": 0.5099731314764989,
+    "brier": 0.335630165723175,
+    "ece": 0.053903984308242794,
+    "mean_conf": 0.7203543915748596,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.75,
+    "cov@0.7": 0.599,
+    "acc@0.7": 0.8430717863105175,
+    "cov@0.9": 0.013,
+    "acc@0.9": 1.0
+  },
+  "wic": {
+    "n": 638,
+    "acc": 0.5909090909090909,
+    "nll": 0.7010703101999811,
+    "brier": 0.49889730712479824,
+    "ece": 0.10223101868898521,
+    "mean_conf": 0.6885135976311555,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.5909090909090909,
+    "cov@0.7": 0.45768025078369906,
+    "acc@0.7": 0.6541095890410958,
+    "cov@0.9": 0.0219435736677116,
+    "acc@0.9": 0.7142857142857143
+  },
+  "subj": {
+    "n": 1000,
+    "acc": 0.683,
+    "nll": 0.5926543222501754,
+    "brier": 0.41087208590491625,
+    "ece": 0.12136625003814697,
+    "mean_conf": 0.7989181578159332,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.683,
+    "cov@0.7": 0.733,
+    "acc@0.7": 0.7517053206002728,
+    "cov@0.9": 0.31,
+    "acc@0.9": 0.9193548387096774
+  },
+  "spam": {
+    "n": 1000,
+    "acc": 0.748,
+    "nll": 0.47593394568919184,
+    "brier": 0.32194300022971223,
+    "ece": 0.10650516629219058,
+    "mean_conf": 0.8214155542850494,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.748,
+    "cov@0.7": 0.823,
+    "acc@0.7": 0.8031591737545565,
+    "cov@0.9": 0.359,
+    "acc@0.9": 0.9721448467966574
+  },
+  "counterfactual": {
+    "n": 1000,
+    "acc": 0.818,
+    "nll": 0.4552239557499372,
+    "brier": 0.283997494798605,
+    "ece": 0.1028595191836357,
+    "mean_conf": 0.7230052413344383,
+    "cov@0.5": 1.0,
+    "acc@0.5": 0.818,
+    "cov@0.7": 0.626,
+    "acc@0.7": 0.9137380191693291,
+    "cov@0.9": 0.012,
+    "acc@0.9": 0.8333333333333334
+  },
+  "cb": {
+    "n": 56,
+    "acc": 0.875,
+    "nll": 0.38689621339204633,
+    "brier": 0.1985667895578788,
+    "ece": 0.12361876879419596,
+    "mean_conf": 0.7997828998735973,
+    "cov@0.5": 0.9821428571428571,
+    "acc@0.5": 0.8909090909090909,
+    "cov@0.7": 0.75,
+    "acc@0.7": 0.9761904761904762,
+    "cov@0.9": 0.2857142857142857,
+    "acc@0.9": 1.0
+  },
+  "arc_challenge": {
+    "n": 1000,
+    "acc": 0.788,
+    "nll": 0.5769762419571128,
+    "brier": 0.3077434279537125,
+    "ece": 0.037862662196159365,
+    "mean_conf": 0.8004423229694366,
+    "cov@0.5": 0.892,
+    "acc@0.5": 0.827354260089686,
+    "cov@0.7": 0.711,
+    "acc@0.7": 0.8874824191279888,
+    "cov@0.9": 0.443,
+    "acc@0.9": 0.9616252821670429
+  },
+  "stance_climate": {
+    "n": 169,
+    "acc": 0.7100591715976331,
+    "nll": 0.7974690104995089,
+    "brier": 0.43555373727561586,
+    "ece": 0.081085235938518,
+    "mean_conf": 0.7026936018608025,
+    "cov@0.5": 0.8579881656804734,
+    "acc@0.5": 0.7241379310344828,
+    "cov@0.7": 0.5029585798816568,
+    "acc@0.7": 0.8941176470588236,
+    "cov@0.9": 0.11834319526627218,
+    "acc@0.9": 0.95
+  },
+  "trec": {
+    "n": 500,
+    "acc": 0.728,
+    "nll": 0.7760585628620453,
+    "brier": 0.3906730053560261,
+    "ece": 0.044884639918804176,
+    "mean_conf": 0.7680628853440284,
+    "cov@0.5": 0.886,
+    "acc@0.5": 0.781038374717833,
+    "cov@0.7": 0.66,
+    "acc@0.7": 0.8333333333333334,
+    "cov@0.9": 0.348,
+    "acc@0.9": 0.8908045977011494
+  },
+  "sst5": {
+    "n": 1000,
+    "acc": 0.57,
+    "nll": 0.9575963019793726,
+    "brier": 0.5502522678043481,
+    "ece": 0.04297695386409761,
+    "mean_conf": 0.5516400979757309,
+    "score_mae": 0.5123533978683991,
+    "cov@0.5": 0.662,
+    "acc@0.5": 0.6314199395770392,
+    "cov@0.7": 0.108,
+    "acc@0.7": 0.7037037037037037,
+    "cov@0.9": 0.003,
+    "acc@0.9": 0.6666666666666666
+  },
+  "fin_sentiment": {
+    "n": 1000,
+    "acc": 0.821,
+    "nll": 0.4648561801968352,
+    "brier": 0.26997927134112937,
+    "ece": 0.07628561544418336,
+    "mean_conf": 0.7507960308790207,
+    "cov@0.5": 0.981,
+    "acc@0.5": 0.8297655453618756,
+    "cov@0.7": 0.669,
+    "acc@0.7": 0.9118086696562033,
+    "cov@0.9": 0.11,
+    "acc@0.9": 0.9727272727272728
+  },
+  "arc_easy": {
+    "n": 1000,
+    "acc": 0.891,
+    "nll": 0.3075507787046519,
+    "brier": 0.1610970984803839,
+    "ece": 0.01694161868095395,
+    "mean_conf": 0.8852402422428131,
+    "cov@0.5": 0.956,
+    "acc@0.5": 0.9131799163179917,
+    "cov@0.7": 0.858,
+    "acc@0.7": 0.9463869463869464,
+    "cov@0.9": 0.691,
+    "acc@0.9": 0.9768451519536903
+  },
+  "newsgroups": {
+    "n": 1000,
+    "acc": 0.646,
+    "nll": 1.2419649776122939,
+    "brier": 0.46508339674462584,
+    "ece": 0.05238002217561006,
+    "mean_conf": 0.6502969245538115,
+    "cov@0.5": 0.676,
+    "acc@0.5": 0.834319526627219,
+    "cov@0.7": 0.492,
+    "acc@0.7": 0.9004065040650406,
+    "cov@0.9": 0.232,
+    "acc@0.9": 0.9870689655172413
+  }
+}
\ No newline at end of file
diff --git a/v0007/manifest.json b/v0007/manifest.json
new file mode 100644
index 0000000000000000000000000000000000000000..6d06e8699c795d8d5faa7a7d7289b643d02c86ce
--- /dev/null
+++ b/v0007/manifest.json
@@ -0,0 +1,1337 @@
+{
+  "version": "v0007",
+  "parent": null,
+  "kind": "lmhead",
+  "backbone": "Qwen/Qwen2.5-1.5B-Instruct",
+  "max_len": 1024,
+  "created": "2026-09-17T15:58:07+00:00",
+  "tasks": [
+    "ag_news",
+    "banking77",
+    "boolq",
+    "clinc150",
+    "commonsense_qa",
+    "copa",
+    "dbpedia",
+    "emotion",
+    "facts",
+    "fits",
+    "imdb",
+    "jailbreak",
+    "massive_intent",
+    "mnli",
+    "mrpc",
+    "openbookqa",
+    "paws",
+    "qnli",
+    "read",
+    "rte",
+    "sciq",
+    "sst2",
+    "swag",
+    "tweet_emoji",
+    "tweet_hate",
+    "tweet_irony",
+    "tweet_offensive",
+    "tweet_sentiment",
+    "yahoo",
+    "yelp",
+    "anli",
+    "winogrande",
+    "hellaswag",
+    "race",
+    "scitail",
+    "qqp",
+    "stsb",
+    "toxic",
+    "stance_abortion",
+    "stance_atheism",
+    "stance_feminist",
+    "stance_hillary",
+    "match",
+    "goemo_soft",
+    "reason",
+    "formality",
+    "politeness",
+    "strategyqa",
+    "vitaminc",
+    "ruletaker",
+    "proofwriter",
+    "folio",
+    "logiqa",
+    "tracie",
+    "temporal_nli",
+    "piqa",
+    "siqa",
+    "clutrr",
+    "gsm8k",
+    "svamp",
+    "aqua"
+  ],
+  "trained_on": [
+    "ag_news",
+    "banking77",
+    "boolq",
+    "clinc150",
+    "commonsense_qa",
+    "copa",
+    "dbpedia",
+    "emotion",
+    "facts",
+    "fits",
+    "imdb",
+    "jailbreak",
+    "massive_intent",
+    "mnli",
+    "mrpc",
+    "openbookqa",
+    "paws",
+    "qnli",
+    "read",
+    "rte",
+    "sciq",
+    "sst2",
+    "swag",
+    "tweet_emoji",
+    "tweet_hate",
+    "tweet_irony",
+    "tweet_offensive",
+    "tweet_sentiment",
+    "yahoo",
+    "yelp",
+    "anli",
+    "winogrande",
+    "hellaswag",
+    "race",
+    "scitail",
+    "qqp",
+    "stsb",
+    "toxic",
+    "stance_abortion",
+    "stance_atheism",
+    "stance_feminist",
+    "stance_hillary",
+    "match",
+    "goemo_soft",
+    "reason",
+    "formality",
+    "politeness",
+    "strategyqa",
+    "vitaminc",
+    "ruletaker",
+    "proofwriter",
+    "folio",
+    "logiqa",
+    "tracie",
+    "temporal_nli",
+    "piqa",
+    "siqa",
+    "clutrr",
+    "gsm8k",
+    "svamp",
+    "aqua"
+  ],
+  "holdout": [
+    "probe",
+    "bbh",
+    "cola",
+    "wic",
+    "subj",
+    "spam",
+    "counterfactual",
+    "cb",
+    "arc_challenge",
+    "stance_climate",
+    "trec",
+    "sst5",
+    "fin_sentiment",
+    "arc_easy",
+    "newsgroups"
+  ],
+  "steps": 12118,
+  "train_examples": 158246,
+  "args": {
+    "cmd": "lmtrain",
+    "lora_r": 16,
+    "loss": "mix",
+    "max_per_task": 3000,
+    "epochs": 1,
+    "lr": 0.0001,
+    "anchor": 0.1
+  },
+  "metrics": {
+    "ag_news": {
+      "n": 300,
+      "acc": 0.9233333333333333,
+      "nll": 0.23956205062784391,
+      "brier": 0.12588630912802798,
+      "ece": 0.03565445333719247,
+      "mean_conf": 0.9105748584866524,
+      "cov@0.5": 0.9833333333333333,
+      "acc@0.5": 0.9288135593220339,
+      "cov@0.7": 0.88,
+      "acc@0.7": 0.9545454545454546,
+      "cov@0.9": 0.77,
+      "acc@0.9": 0.9653679653679653
+    },
+    "boolq": {
+      "n": 300,
+      "acc": 0.8766666666666667,
+      "nll": 0.33278360864483053,
+      "brier": 0.19905847383131203,
+      "ece": 0.048838321963946066,
+      "mean_conf": 0.8610335459311803,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.8766666666666667,
+      "cov@0.7": 0.8866666666666667,
+      "acc@0.7": 0.9060150375939849,
+      "cov@0.9": 0.4666666666666667,
+      "acc@0.9": 0.9642857142857143
+    },
+    "commonsense_qa": {
+      "n": 300,
+      "acc": 0.7866666666666666,
+      "nll": 0.5812667453454432,
+      "brier": 0.2960317921716136,
+      "ece": 0.05657371540864312,
+      "mean_conf": 0.7944276158014933,
+      "cov@0.5": 0.8933333333333333,
+      "acc@0.5": 0.832089552238806,
+      "cov@0.7": 0.7,
+      "acc@0.7": 0.9047619047619048,
+      "cov@0.9": 0.42333333333333334,
+      "acc@0.9": 0.952755905511811
+    },
+    "copa": {
+      "n": 200,
+      "acc": 0.93,
+      "nll": 0.17489093500636219,
+      "brier": 0.10265215895968784,
+      "ece": 0.03722041517496112,
+      "mean_conf": 0.8977384361624717,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.93,
+      "cov@0.7": 0.9,
+      "acc@0.7": 0.9611111111111111,
+      "cov@0.9": 0.69,
+      "acc@0.9": 1.0
+    },
+    "dbpedia": {
+      "n": 300,
+      "acc": 0.97,
+      "nll": 0.12029726431718998,
+      "brier": 0.0467073057529959,
+      "ece": 0.011912721196810366,
+      "mean_conf": 0.9797261367241542,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.97,
+      "cov@0.7": 0.9866666666666667,
+      "acc@0.7": 0.9797297297297297,
+      "cov@0.9": 0.9766666666666667,
+      "acc@0.9": 0.9829351535836177
+    },
+    "emotion": {
+      "n": 300,
+      "acc": 0.7766666666666666,
+      "nll": 0.6417595775959655,
+      "brier": 0.31279347659514983,
+      "ece": 0.07161758591731388,
+      "mean_conf": 0.8046423467993736,
+      "cov@0.5": 0.92,
+      "acc@0.5": 0.8043478260869565,
+      "cov@0.7": 0.7233333333333334,
+      "acc@0.7": 0.9032258064516129,
+      "cov@0.9": 0.4266666666666667,
+      "acc@0.9": 0.9609375
+    },
+    "facts": {
+      "n": 54,
+      "acc": 0.9629629629629629,
+      "nll": 0.10233185567308967,
+      "brier": 0.0580309563229181,
+      "ece": 0.03901768503365692,
+      "mean_conf": 0.985795874286581,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.9629629629629629,
+      "cov@0.7": 1.0,
+      "acc@0.7": 0.9629629629629629,
+      "cov@0.9": 0.9444444444444444,
+      "acc@0.9": 1.0
+    },
+    "fits": {
+      "n": 214,
+      "acc": 0.8598130841121495,
+      "nll": 0.4789074549575723,
+      "brier": 0.09971697154134483,
+      "ece": 0.08341867177285883,
+      "mean_conf": 0.818978947456752,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.8598130841121495,
+      "cov@0.7": 0.8037383177570093,
+      "acc@0.7": 0.936046511627907,
+      "cov@0.9": 0.3317757009345794,
+      "acc@0.9": 1.0
+    },
+    "imdb": {
+      "n": 300,
+      "acc": 0.95,
+      "nll": 0.1449115194526972,
+      "brier": 0.07315853850730186,
+      "ece": 0.028431474765141816,
+      "mean_conf": 0.9752282838026682,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.95,
+      "cov@0.7": 0.9833333333333333,
+      "acc@0.7": 0.9627118644067797,
+      "cov@0.9": 0.95,
+      "acc@0.9": 0.9754385964912281
+    },
+    "jailbreak": {
+      "n": 200,
+      "acc": 0.98,
+      "nll": 0.07703215111456595,
+      "brier": 0.03552832502947578,
+      "ece": 0.014975480735301977,
+      "mean_conf": 0.972049820125103,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.98,
+      "cov@0.7": 0.995,
+      "acc@0.7": 0.9798994974874372,
+      "cov@0.9": 0.965,
+      "acc@0.9": 0.9896373056994818
+    },
+    "mnli": {
+      "n": 300,
+      "acc": 0.8466666666666667,
+      "nll": 0.44322894987837685,
+      "brier": 0.24040505319501904,
+      "ece": 0.05213867117961251,
+      "mean_conf": 0.8880125307043394,
+      "cov@0.5": 0.9933333333333333,
+      "acc@0.5": 0.8489932885906041,
+      "cov@0.7": 0.9,
+      "acc@0.7": 0.8814814814814815,
+      "cov@0.9": 0.64,
+      "acc@0.9": 0.9427083333333334
+    },
+    "mrpc": {
+      "n": 200,
+      "acc": 0.83,
+      "nll": 0.3825738710855869,
+      "brier": 0.2414001594036636,
+      "ece": 0.0455596360564232,
+      "mean_conf": 0.8177212104201317,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.83,
+      "cov@0.7": 0.745,
+      "acc@0.7": 0.9060402684563759,
+      "cov@0.9": 0.415,
+      "acc@0.9": 0.9518072289156626
+    },
+    "openbookqa": {
+      "n": 300,
+      "acc": 0.8166666666666667,
+      "nll": 0.4884890429825752,
+      "brier": 0.26006302091594946,
+      "ece": 0.06033645391464233,
+      "mean_conf": 0.8382209448019663,
+      "cov@0.5": 0.9266666666666666,
+      "acc@0.5": 0.8525179856115108,
+      "cov@0.7": 0.7966666666666666,
+      "acc@0.7": 0.899581589958159,
+      "cov@0.9": 0.52,
+      "acc@0.9": 0.9743589743589743
+    },
+    "paws": {
+      "n": 300,
+      "acc": 0.93,
+      "nll": 0.2171823905014709,
+      "brier": 0.11854569746905995,
+      "ece": 0.028286268909772223,
+      "mean_conf": 0.9089529289801915,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.93,
+      "cov@0.7": 0.9166666666666666,
+      "acc@0.7": 0.9490909090909091,
+      "cov@0.9": 0.7366666666666667,
+      "acc@0.9": 0.9638009049773756
+    },
+    "qnli": {
+      "n": 300,
+      "acc": 0.88,
+      "nll": 0.29461164414824,
+      "brier": 0.1746824483593742,
+      "ece": 0.03399742662906641,
+      "mean_conf": 0.9018590325117111,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.88,
+      "cov@0.7": 0.92,
+      "acc@0.7": 0.9166666666666666,
+      "cov@0.9": 0.6933333333333334,
+      "acc@0.9": 0.9423076923076923
+    },
+    "read": {
+      "n": 300,
+      "acc": 1.0,
+      "nll": 0.013479419595217349,
+      "brier": 0.0006464961621803141,
+      "ece": 0.01327262739340459,
+      "mean_conf": 0.9867273726065954,
+      "cov@0.5": 1.0,
+      "acc@0.5": 1.0,
+      "cov@0.7": 1.0,
+      "acc@0.7": 1.0,
+      "cov@0.9": 0.9966666666666667,
+      "acc@0.9": 1.0
+    },
+    "rte": {
+      "n": 200,
+      "acc": 0.875,
+      "nll": 0.2619901953560894,
+      "brier": 0.1655817686780098,
+      "ece": 0.07393659174442292,
+      "mean_conf": 0.8983567506074905,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.875,
+      "cov@0.7": 0.915,
+      "acc@0.7": 0.912568306010929,
+      "cov@0.9": 0.69,
+      "acc@0.9": 0.9855072463768116
+    },
+    "sciq": {
+      "n": 300,
+      "acc": 0.9833333333333333,
+      "nll": 0.0542967460035165,
+      "brier": 0.02836314450778487,
+      "ece": 0.013889081676801088,
+      "mean_conf": 0.9804451249043147,
+      "cov@0.5": 0.9966666666666667,
+      "acc@0.5": 0.9866220735785953,
+      "cov@0.7": 0.99,
+      "acc@0.7": 0.9865319865319865,
+      "cov@0.9": 0.95,
+      "acc@0.9": 0.9929824561403509
+    },
+    "sst2": {
+      "n": 300,
+      "acc": 0.94,
+      "nll": 0.15217635479920968,
+      "brier": 0.08669829778686115,
+      "ece": 0.02306454201539363,
+      "mean_conf": 0.9453726333379745,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.94,
+      "cov@0.7": 0.9633333333333334,
+      "acc@0.7": 0.9584775086505191,
+      "cov@0.9": 0.82,
+      "acc@0.9": 0.991869918699187
+    },
+    "swag": {
+      "n": 300,
+      "acc": 0.7666666666666667,
+      "nll": 0.7234309196216375,
+      "brier": 0.3695786167661231,
+      "ece": 0.06823263516028721,
+      "mean_conf": 0.7794082881013552,
+      "cov@0.5": 0.91,
+      "acc@0.5": 0.7875457875457875,
+      "cov@0.7": 0.69,
+      "acc@0.7": 0.8405797101449275,
+      "cov@0.9": 0.3333333333333333,
+      "acc@0.9": 0.9
+    },
+    "tweet_emoji": {
+      "n": 300,
+      "acc": 0.24333333333333335,
+      "nll": 2.5063694445877984,
+      "brier": 0.83829218239292,
+      "ece": 0.08510932529966037,
+      "mean_conf": 0.2803319871922334,
+      "cov@0.5": 0.12,
+      "acc@0.5": 0.8055555555555556,
+      "cov@0.7": 0.07666666666666666,
+      "acc@0.7": 0.9130434782608695,
+      "cov@0.9": 0.0033333333333333335,
+      "acc@0.9": 1.0
+    },
+    "tweet_hate": {
+      "n": 300,
+      "acc": 0.7166666666666667,
+      "nll": 0.5309249597827629,
+      "brier": 0.36006795343339115,
+      "ece": 0.09840704739093784,
+      "mean_conf": 0.8124431739250819,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.7166666666666667,
+      "cov@0.7": 0.7866666666666666,
+      "acc@0.7": 0.7966101694915254,
+      "cov@0.9": 0.32666666666666666,
+      "acc@0.9": 0.9285714285714286
+    },
+    "tweet_irony": {
+      "n": 300,
+      "acc": 0.71,
+      "nll": 0.5534378644031082,
+      "brier": 0.37644439644653505,
+      "ece": 0.03676844378312428,
+      "mean_conf": 0.7351771769920985,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.71,
+      "cov@0.7": 0.5866666666666667,
+      "acc@0.7": 0.8068181818181818,
+      "cov@0.9": 0.12333333333333334,
+      "acc@0.9": 0.972972972972973
+    },
+    "tweet_offensive": {
+      "n": 300,
+      "acc": 0.7866666666666666,
+      "nll": 0.47620018437820133,
+      "brier": 0.3128923872633037,
+      "ece": 0.05815195361773175,
+      "mean_conf": 0.8035398570696513,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.7866666666666666,
+      "cov@0.7": 0.76,
+      "acc@0.7": 0.8421052631578947,
+      "cov@0.9": 0.3333333333333333,
+      "acc@0.9": 0.93
+    },
+    "tweet_sentiment": {
+      "n": 300,
+      "acc": 0.7266666666666667,
+      "nll": 0.6050564579947201,
+      "brier": 0.36215817406899686,
+      "ece": 0.07843278278907143,
+      "mean_conf": 0.7284325797359149,
+      "score_mae": 0.3644068883561219,
+      "cov@0.5": 0.96,
+      "acc@0.5": 0.7291666666666666,
+      "cov@0.7": 0.5333333333333333,
+      "acc@0.7": 0.875,
+      "cov@0.9": 0.16666666666666666,
+      "acc@0.9": 0.92
+    },
+    "yahoo": {
+      "n": 300,
+      "acc": 0.7166666666666667,
+      "nll": 0.8738976731967112,
+      "brier": 0.3950192633049044,
+      "ece": 0.092816769828399,
+      "mean_conf": 0.8022240548829238,
+      "cov@0.5": 0.9133333333333333,
+      "acc@0.5": 0.7627737226277372,
+      "cov@0.7": 0.7466666666666667,
+      "acc@0.7": 0.8348214285714286,
+      "cov@0.9": 0.4166666666666667,
+      "acc@0.9": 0.92
+    },
+    "yelp": {
+      "n": 300,
+      "acc": 0.71,
+      "nll": 0.731893961859198,
+      "brier": 0.42142868253511284,
+      "ece": 0.09561944127082822,
+      "mean_conf": 0.7473884936173757,
+      "score_mae": 0.3743417001541093,
+      "cov@0.5": 0.9533333333333334,
+      "acc@0.5": 0.7237762237762237,
+      "cov@0.7": 0.6266666666666667,
+      "acc@0.7": 0.7712765957446809,
+      "cov@0.9": 0.19,
+      "acc@0.9": 0.9649122807017544
+    },
+    "anli": {
+      "n": 300,
+      "acc": 0.5366666666666666,
+      "nll": 1.1646257930212216,
+      "brier": 0.675594648318733,
+      "ece": 0.223523634771506,
+      "mean_conf": 0.7592115387320518,
+      "cov@0.5": 0.9433333333333334,
+      "acc@0.5": 0.5547703180212014,
+      "cov@0.7": 0.65,
+      "acc@0.7": 0.558974358974359,
+      "cov@0.9": 0.23,
+      "acc@0.9": 0.5507246376811594
+    },
+    "winogrande": {
+      "n": 300,
+      "acc": 0.7966666666666666,
+      "nll": 0.46445836318766226,
+      "brier": 0.3013142673360519,
+      "ece": 0.07452347179253899,
+      "mean_conf": 0.8619708905617396,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.7966666666666666,
+      "cov@0.7": 0.87,
+      "acc@0.7": 0.8275862068965517,
+      "cov@0.9": 0.55,
+      "acc@0.9": 0.9030303030303031
+    },
+    "hellaswag": {
+      "n": 300,
+      "acc": 0.8566666666666667,
+      "nll": 0.3762581950318828,
+      "brier": 0.19513976820616458,
+      "ece": 0.05684188375870386,
+      "mean_conf": 0.8465098922451337,
+      "cov@0.5": 0.9366666666666666,
+      "acc@0.5": 0.9074733096085409,
+      "cov@0.7": 0.7833333333333333,
+      "acc@0.7": 0.9446808510638298,
+      "cov@0.9": 0.5533333333333333,
+      "acc@0.9": 0.9819277108433735
+    },
+    "race": {
+      "n": 300,
+      "acc": 0.7933333333333333,
+      "nll": 0.5311142158904583,
+      "brier": 0.2739426106933522,
+      "ece": 0.0815848172704379,
+      "mean_conf": 0.8720341417193412,
+      "cov@0.5": 0.93,
+      "acc@0.5": 0.8422939068100358,
+      "cov@0.7": 0.8466666666666667,
+      "acc@0.7": 0.8818897637795275,
+      "cov@0.9": 0.6566666666666666,
+      "acc@0.9": 0.934010152284264
+    },
+    "scitail": {
+      "n": 300,
+      "acc": 0.97,
+      "nll": 0.1026227491797548,
+      "brier": 0.05202258655692825,
+      "ece": 0.028236998518308055,
+      "mean_conf": 0.950906420747439,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.97,
+      "cov@0.7": 0.9833333333333333,
+      "acc@0.7": 0.9728813559322034,
+      "cov@0.9": 0.8466666666666667,
+      "acc@0.9": 0.9921259842519685
+    },
+    "qqp": {
+      "n": 300,
+      "acc": 0.88,
+      "nll": 0.2769447782152307,
+      "brier": 0.17137568112639565,
+      "ece": 0.045283984939257296,
+      "mean_conf": 0.8734472642342249,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.88,
+      "cov@0.7": 0.8533333333333334,
+      "acc@0.7": 0.93359375,
+      "cov@0.9": 0.5766666666666667,
+      "acc@0.9": 0.9710982658959537
+    },
+    "stsb": {
+      "n": 287,
+      "acc": 0.6167247386759582,
+      "nll": 1.0364299467274243,
+      "brier": 0.2999134000942182,
+      "ece": 0.07439108956150894,
+      "mean_conf": 0.5469927265461314,
+      "score_mae": 0.5417481492766754,
+      "cov@0.5": 0.6306620209059234,
+      "acc@0.5": 0.6850828729281768,
+      "cov@0.7": 0.09407665505226481,
+      "acc@0.7": 0.8518518518518519,
+      "cov@0.9": 0.0,
+      "acc@0.9": NaN
+    },
+    "toxic": {
+      "n": 300,
+      "acc": 0.84,
+      "nll": 0.34017511751074586,
+      "brier": 0.21106674918538404,
+      "ece": 0.04502722958723705,
+      "mean_conf": 0.8586645072698593,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.84,
+      "cov@0.7": 0.88,
+      "acc@0.7": 0.8977272727272727,
+      "cov@0.9": 0.52,
+      "acc@0.9": 0.9615384615384616
+    },
+    "stance_abortion": {
+      "n": 66,
+      "acc": 0.8636363636363636,
+      "nll": 0.44344493585724276,
+      "brier": 0.23815874340787538,
+      "ece": 0.16279316354881634,
+      "mean_conf": 0.7459786872972142,
+      "cov@0.5": 0.8636363636363636,
+      "acc@0.5": 0.8947368421052632,
+      "cov@0.7": 0.5909090909090909,
+      "acc@0.7": 0.9230769230769231,
+      "cov@0.9": 0.22727272727272727,
+      "acc@0.9": 1.0
+    },
+    "stance_atheism": {
+      "n": 52,
+      "acc": 0.7115384615384616,
+      "nll": 0.6112422593374254,
+      "brier": 0.37005177211571516,
+      "ece": 0.12039663585332726,
+      "mean_conf": 0.7864993226069671,
+      "cov@0.5": 0.9423076923076923,
+      "acc@0.5": 0.7346938775510204,
+      "cov@0.7": 0.6538461538461539,
+      "acc@0.7": 0.8235294117647058,
+      "cov@0.9": 0.3269230769230769,
+      "acc@0.9": 1.0
+    },
+    "stance_feminist": {
+      "n": 67,
+      "acc": 0.7313432835820896,
+      "nll": 0.7303124558726009,
+      "brier": 0.4342147904610734,
+      "ece": 0.1691147155726134,
+      "mean_conf": 0.7381730840277316,
+      "cov@0.5": 0.9402985074626866,
+      "acc@0.5": 0.746031746031746,
+      "cov@0.7": 0.5522388059701493,
+      "acc@0.7": 0.7027027027027027,
+      "cov@0.9": 0.19402985074626866,
+      "acc@0.9": 0.8461538461538461
+    },
+    "stance_hillary": {
+      "n": 69,
+      "acc": 0.7246376811594203,
+      "nll": 0.6620274924672167,
+      "brier": 0.3935280964850548,
+      "ece": 0.09731849552928537,
+      "mean_conf": 0.7686952758526456,
+      "cov@0.5": 0.9565217391304348,
+      "acc@0.5": 0.7424242424242424,
+      "cov@0.7": 0.6666666666666666,
+      "acc@0.7": 0.8043478260869565,
+      "cov@0.9": 0.18840579710144928,
+      "acc@0.9": 0.7692307692307693
+    },
+    "match": {
+      "n": 300,
+      "acc": 0.99,
+      "nll": 0.03724694484885282,
+      "brier": 0.016051036461267973,
+      "ece": 0.009268070856730138,
+      "mean_conf": 0.9877331558863321,
+      "cov@0.5": 0.9966666666666667,
+      "acc@0.5": 0.9933110367892977,
+      "cov@0.7": 0.9966666666666667,
+      "acc@0.7": 0.9933110367892977,
+      "cov@0.9": 0.99,
+      "acc@0.9": 0.9932659932659933
+    },
+    "reason": {
+      "n": 300,
+      "acc": 0.9466666666666667,
+      "nll": 0.17577242359452994,
+      "brier": 0.09125921721706085,
+      "ece": 0.04738917231559757,
+      "mean_conf": 0.9024230941136678,
+      "cov@0.5": 0.9833333333333333,
+      "acc@0.5": 0.9559322033898305,
+      "cov@0.7": 0.88,
+      "acc@0.7": 0.9772727272727273,
+      "cov@0.9": 0.7733333333333333,
+      "acc@0.9": 0.9913793103448276
+    },
+    "formality": {
+      "n": 300,
+      "acc": 0.6066666666666667,
+      "nll": 0.9962800608084179,
+      "brier": 0.25308033830347887,
+      "ece": 0.07329747378826143,
+      "mean_conf": 0.5359433833758036,
+      "score_mae": 0.47120167226336584,
+      "cov@0.5": 0.6866666666666666,
+      "acc@0.5": 0.6699029126213593,
+      "cov@0.7": 0.0033333333333333335,
+      "acc@0.7": 1.0,
+      "cov@0.9": 0.0,
+      "acc@0.9": NaN
+    },
+    "politeness": {
+      "n": 300,
+      "acc": 0.8633333333333333,
+      "nll": 0.3753622173094621,
+      "brier": 0.18929635618156296,
+      "ece": 0.026931450863679252,
+      "mean_conf": 0.8615820496280988,
+      "score_mae": 0.21370556724568207,
+      "cov@0.5": 0.9633333333333334,
+      "acc@0.5": 0.8858131487889274,
+      "cov@0.7": 0.8266666666666667,
+      "acc@0.7": 0.9435483870967742,
+      "cov@0.9": 0.5766666666666667,
+      "acc@0.9": 0.9826589595375722
+    },
+    "strategyqa": {
+      "n": 200,
+      "acc": 0.67,
+      "nll": 0.5862248309774405,
+      "brier": 0.40554963519042464,
+      "ece": 0.05940887540578843,
+      "mean_conf": 0.6801840284466744,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.67,
+      "cov@0.7": 0.385,
+      "acc@0.7": 0.8051948051948052,
+      "cov@0.9": 0.07,
+      "acc@0.9": 1.0
+    },
+    "vitaminc": {
+      "n": 300,
+      "acc": 0.82,
+      "nll": 0.5214132864032636,
+      "brier": 0.2848927857191196,
+      "ece": 0.03640172024567924,
+      "mean_conf": 0.8344497634967168,
+      "cov@0.5": 0.9666666666666667,
+      "acc@0.5": 0.8275862068965517,
+      "cov@0.7": 0.83,
+      "acc@0.7": 0.8634538152610441,
+      "cov@0.9": 0.4533333333333333,
+      "acc@0.9": 0.9338235294117647
+    },
+    "ruletaker": {
+      "n": 300,
+      "acc": 0.8066666666666666,
+      "nll": 0.3848169850830573,
+      "brier": 0.25008683944182986,
+      "ece": 0.030118134220441215,
+      "mean_conf": 0.812346151471138,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.8066666666666666,
+      "cov@0.7": 0.7133333333333334,
+      "acc@0.7": 0.9018691588785047,
+      "cov@0.9": 0.43666666666666665,
+      "acc@0.9": 0.9618320610687023
+    },
+    "proofwriter": {
+      "n": 300,
+      "acc": 0.7933333333333333,
+      "nll": 0.4490651794863921,
+      "brier": 0.2722325912094693,
+      "ece": 0.06963174422581991,
+      "mean_conf": 0.8091626433531444,
+      "cov@0.5": 0.98,
+      "acc@0.5": 0.7993197278911565,
+      "cov@0.7": 0.7633333333333333,
+      "acc@0.7": 0.8777292576419214,
+      "cov@0.9": 0.37333333333333335,
+      "acc@0.9": 0.9910714285714286
+    },
+    "folio": {
+      "n": 200,
+      "acc": 0.625,
+      "nll": 0.7944451580787752,
+      "brier": 0.4657608423671937,
+      "ece": 0.11621890529990193,
+      "mean_conf": 0.6586561058461666,
+      "cov@0.5": 0.835,
+      "acc@0.5": 0.688622754491018,
+      "cov@0.7": 0.375,
+      "acc@0.7": 0.8533333333333334,
+      "cov@0.9": 0.06,
+      "acc@0.9": 0.75
+    },
+    "logiqa": {
+      "n": 300,
+      "acc": 0.5833333333333334,
+      "nll": 0.6570202462503513,
+      "brier": 0.46499107764281516,
+      "ece": 0.061852243741353355,
+      "mean_conf": 0.6316624116897583,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.5833333333333334,
+      "cov@0.7": 0.23666666666666666,
+      "acc@0.7": 0.7323943661971831,
+      "cov@0.9": 0.013333333333333334,
+      "acc@0.9": 0.75
+    },
+    "tracie": {
+      "n": 200,
+      "acc": 0.77,
+      "nll": 0.5158300579084877,
+      "brier": 0.3394760687221474,
+      "ece": 0.07212192535400391,
+      "mean_conf": 0.70532983481884,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.77,
+      "cov@0.7": 0.49,
+      "acc@0.7": 0.8673469387755102,
+      "cov@0.9": 0.03,
+      "acc@0.9": 1.0
+    },
+    "temporal_nli": {
+      "n": 300,
+      "acc": 0.8633333333333333,
+      "nll": 0.3496475835064699,
+      "brier": 0.20055396242126278,
+      "ece": 0.07610381027062735,
+      "mean_conf": 0.7972539271910986,
+      "cov@0.5": 0.9866666666666667,
+      "acc@0.5": 0.8614864864864865,
+      "cov@0.7": 0.7666666666666667,
+      "acc@0.7": 0.9217391304347826,
+      "cov@0.9": 0.26666666666666666,
+      "acc@0.9": 1.0
+    },
+    "piqa": {
+      "n": 300,
+      "acc": 0.8166666666666667,
+      "nll": 0.36417624160320555,
+      "brier": 0.23331890879921213,
+      "ece": 0.038980930248896296,
+      "mean_conf": 0.8269510519504547,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.8166666666666667,
+      "cov@0.7": 0.7633333333333333,
+      "acc@0.7": 0.9126637554585153,
+      "cov@0.9": 0.43666666666666665,
+      "acc@0.9": 0.9618320610687023
+    },
+    "siqa": {
+      "n": 300,
+      "acc": 0.8233333333333334,
+      "nll": 0.4337533873205736,
+      "brier": 0.24543190225066103,
+      "ece": 0.034700063069661474,
+      "mean_conf": 0.8447816316286723,
+      "cov@0.5": 0.9633333333333334,
+      "acc@0.5": 0.8408304498269896,
+      "cov@0.7": 0.7833333333333333,
+      "acc@0.7": 0.9148936170212766,
+      "cov@0.9": 0.5533333333333333,
+      "acc@0.9": 0.9457831325301205
+    },
+    "clutrr": {
+      "n": 300,
+      "acc": 0.6133333333333333,
+      "nll": 0.9127365656335344,
+      "brier": 0.4770832175168861,
+      "ece": 0.04643365234136578,
+      "mean_conf": 0.5962887792785962,
+      "cov@0.5": 0.72,
+      "acc@0.5": 0.6944444444444444,
+      "cov@0.7": 0.21333333333333335,
+      "acc@0.7": 0.9375,
+      "cov@0.9": 0.08333333333333333,
+      "acc@0.9": 1.0
+    },
+    "gsm8k": {
+      "n": 300,
+      "acc": 0.74,
+      "nll": 0.589682709113467,
+      "brier": 0.32688956856451207,
+      "ece": 0.06247660587231318,
+      "mean_conf": 0.7070875727136929,
+      "cov@0.5": 0.7866666666666666,
+      "acc@0.5": 0.826271186440678,
+      "cov@0.7": 0.53,
+      "acc@0.7": 0.9308176100628931,
+      "cov@0.9": 0.25,
+      "acc@0.9": 1.0
+    },
+    "svamp": {
+      "n": 100,
+      "acc": 0.63,
+      "nll": 0.8085613738920417,
+      "brier": 0.4651299440112998,
+      "ece": 0.12156755775213245,
+      "mean_conf": 0.601172327697277,
+      "cov@0.5": 0.68,
+      "acc@0.5": 0.75,
+      "cov@0.7": 0.32,
+      "acc@0.7": 0.875,
+      "cov@0.9": 0.11,
+      "acc@0.9": 1.0
+    },
+    "aqua": {
+      "n": 253,
+      "acc": 0.383399209486166,
+      "nll": 1.467006593256583,
+      "brier": 0.7354768193410129,
+      "ece": 0.05388568453637978,
+      "mean_conf": 0.3621667535173092,
+      "cov@0.5": 0.09486166007905138,
+      "acc@0.5": 0.6666666666666666,
+      "cov@0.7": 0.019762845849802372,
+      "acc@0.7": 0.4,
+      "cov@0.9": 0.0,
+      "acc@0.9": NaN
+    }
+  },
+  "unseen_test_uncalibrated": {
+    "probe": {
+      "n": 97,
+      "acc": 0.8969072164948454,
+      "nll": 0.20691493444770823,
+      "brier": 0.12394450383288905,
+      "ece": 0.06851052070401381,
+      "mean_conf": 0.9103295477395205,
+      "score_mae": 0.18648642087646294,
+      "cov@0.5": 0.9896907216494846,
+      "acc@0.5": 0.90625,
+      "cov@0.7": 0.9072164948453608,
+      "acc@0.7": 0.9545454545454546,
+      "cov@0.9": 0.7628865979381443,
+      "acc@0.9": 0.9864864864864865,
+      "families": {
+        "desc": [
+          13,
+          15
+        ],
+        "negation": [
+          10,
+          10
+        ],
+        "logic": [
+          14,
+          17
+        ],
+        "score": [
+          10,
+          12
+        ],
+        "json": [
+          6,
+          6
+        ],
+        "taxonomy": [
+          14,
+          14
+        ],
+        "plausible": [
+          4,
+          4
+        ],
+        "twist": [
+          3,
+          3
+        ],
+        "time": [
+          4,
+          5
+        ],
+        "intent": [
+          2,
+          3
+        ],
+        "compare": [
+          3,
+          4
+        ],
+        "criteria": [
+          4,
+          4
+        ]
+      }
+    },
+    "bbh": {
+      "n": 1000,
+      "acc": 0.517,
+      "nll": 1.108523054891869,
+      "brier": 0.6035352775259032,
+      "ece": 0.07363909149914981,
+      "mean_conf": 0.5724366051629186,
+      "cov@0.5": 0.64,
+      "acc@0.5": 0.6140625,
+      "cov@0.7": 0.273,
+      "acc@0.7": 0.6959706959706959,
+      "cov@0.9": 0.071,
+      "acc@0.9": 0.7887323943661971
+    },
+    "cola": {
+      "n": 1000,
+      "acc": 0.75,
+      "nll": 0.5104252819118784,
+      "brier": 0.33587838373719203,
+      "ece": 0.0569284417629242,
+      "mean_conf": 0.7190280594825744,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.75,
+      "cov@0.7": 0.58,
+      "acc@0.7": 0.8448275862068966,
+      "cov@0.9": 0.013,
+      "acc@0.9": 1.0
+    },
+    "wic": {
+      "n": 638,
+      "acc": 0.5909090909090909,
+      "nll": 0.69998100116263,
+      "brier": 0.4982001306496651,
+      "ece": 0.09791882723850147,
+      "mean_conf": 0.6873193471969855,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.5909090909090909,
+      "cov@0.7": 0.44357366771159873,
+      "acc@0.7": 0.6537102473498233,
+      "cov@0.9": 0.0219435736677116,
+      "acc@0.9": 0.7142857142857143
+    },
+    "subj": {
+      "n": 1000,
+      "acc": 0.682,
+      "nll": 0.5917236264696343,
+      "brier": 0.4104362382942676,
+      "ece": 0.1212564522027969,
+      "mean_conf": 0.7977630772590637,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.682,
+      "cov@0.7": 0.733,
+      "acc@0.7": 0.7517053206002728,
+      "cov@0.9": 0.306,
+      "acc@0.9": 0.9215686274509803
+    },
+    "spam": {
+      "n": 1000,
+      "acc": 0.749,
+      "nll": 0.47590900971852373,
+      "brier": 0.32175875624611966,
+      "ece": 0.10708969771862033,
+      "mean_conf": 0.8200223511457443,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.749,
+      "cov@0.7": 0.816,
+      "acc@0.7": 0.803921568627451,
+      "cov@0.9": 0.353,
+      "acc@0.9": 0.9773371104815864
+    },
+    "counterfactual": {
+      "n": 1000,
+      "acc": 0.82,
+      "nll": 0.4564823726332652,
+      "brier": 0.2849392428188435,
+      "ece": 0.10961719477176668,
+      "mean_conf": 0.7216410273313523,
+      "cov@0.5": 1.0,
+      "acc@0.5": 0.82,
+      "cov@0.7": 0.617,
+      "acc@0.7": 0.9141004862236629,
+      "cov@0.9": 0.01,
+      "acc@0.9": 0.8
+    },
+    "cb": {
+      "n": 56,
+      "acc": 0.875,
+      "nll": 0.3821062575477204,
+      "brier": 0.19684515831431662,
+      "ece": 0.11659146206719534,
+      "mean_conf": 0.8077813791377204,
+      "cov@0.5": 0.9821428571428571,
+      "acc@0.5": 0.8909090909090909,
+      "cov@0.7": 0.75,
+      "acc@0.7": 0.9761904761904762,
+      "cov@0.9": 0.35714285714285715,
+      "acc@0.9": 1.0
+    },
+    "arc_challenge": {
+      "n": 1000,
+      "acc": 0.79,
+      "nll": 0.5784666197380147,
+      "brier": 0.3084519084178791,
+      "ece": 0.04053584739565846,
+      "mean_conf": 0.8078768512308597,
+      "cov@0.5": 0.897,
+      "acc@0.5": 0.8249721293199554,
+      "cov@0.7": 0.719,
+      "acc@0.7": 0.8873435326842837,
+      "cov@0.9": 0.46,
+      "acc@0.9": 0.9586956521739131
+    },
+    "stance_climate": {
+      "n": 169,
+      "acc": 0.7100591715976331,
+      "nll": 0.8004923563430824,
+      "brier": 0.43626973420217074,
+      "ece": 0.08912066789068414,
+      "mean_conf": 0.7112431993498605,
+      "cov@0.5": 0.863905325443787,
+      "acc@0.5": 0.726027397260274,
+      "cov@0.7": 0.5502958579881657,
+      "acc@0.7": 0.8709677419354839,
+      "cov@0.9": 0.1301775147928994,
+      "acc@0.9": 0.9545454545454546
+    },
+    "trec": {
+      "n": 500,
+      "acc": 0.728,
+      "nll": 0.7789360883597788,
+      "brier": 0.3915759426435931,
+      "ece": 0.055817239046096784,
+      "mean_conf": 0.7733147183656692,
+      "cov@0.5": 0.89,
+      "acc@0.5": 0.7797752808988764,
+      "cov@0.7": 0.668,
+      "acc@0.7": 0.8323353293413174,
+      "cov@0.9": 0.362,
+      "acc@0.9": 0.8895027624309392
+    },
+    "sst5": {
+      "n": 1000,
+      "acc": 0.569,
+      "nll": 0.9569207478808627,
+      "brier": 0.5500160496510278,
+      "ece": 0.051933742076158515,
+      "mean_conf": 0.5590169258415699,
+      "score_mae": 0.5127075865020743,
+      "cov@0.5": 0.684,
+      "acc@0.5": 0.6198830409356725,
+      "cov@0.7": 0.117,
+      "acc@0.7": 0.7094017094017094,
+      "cov@0.9": 0.005,
+      "acc@0.9": 0.6
+    },
+    "fin_sentiment": {
+      "n": 1000,
+      "acc": 0.821,
+      "nll": 0.45971791800530354,
+      "brier": 0.26788516663028156,
+      "ece": 0.06331040862202642,
+      "mean_conf": 0.7591086620986461,
+      "cov@0.5": 0.981,
+      "acc@0.5": 0.8287461773700305,
+      "cov@0.7": 0.684,
+      "acc@0.7": 0.9078947368421053,
+      "cov@0.9": 0.139,
+      "acc@0.9": 0.9640287769784173
+    },
+    "arc_easy": {
+      "n": 1000,
+      "acc": 0.893,
+      "nll": 0.30681700111668636,
+      "brier": 0.16114664739945947,
+      "ece": 0.017386740416288377,
+      "mean_conf": 0.8903764767348766,
+      "cov@0.5": 0.959,
+      "acc@0.5": 0.9113660062565172,
+      "cov@0.7": 0.864,
+      "acc@0.7": 0.9444444444444444,
+      "cov@0.9": 0.704,
+      "acc@0.9": 0.9758522727272727
+    },
+    "newsgroups": {
+      "n": 1000,
+      "acc": 0.645,
+      "nll": 1.2438821946531702,
+      "brier": 0.4654980387667391,
+      "ece": 0.0510300065651536,
+      "mean_conf": 0.6572305353507399,
+      "cov@0.5": 0.683,
+      "acc@0.5": 0.8272327964860908,
+      "cov@0.7": 0.5,
+      "acc@0.7": 0.9,
+      "cov@0.9": 0.246,
+      "acc@0.9": 0.9878048780487805
+    }
+  },
+  "summary": {
+    "mean_acc": 0.7922933763477235,
+    "mean_ece": 0.06318428710662419
+  },
+  "history": [],
+  "temperature": 1.0194810628890991,
+  "temperature_by_k": {
+    "2-2": 0.9909,
+    "3-5": 1.0355,
+    "6-20": 1.0193
+  },
+  "calibration": {
+    "before": {
+      "n": 10608,
+      "acc": 0.7931749622926093,
+      "nll": 0.5081595549816602,
+      "brier": 0.26480504720767034,
+      "ece": 0.010474447911776612,
+      "mean_conf": 0.8014159178505982,
+      "score_mae": 0.3837145484503708,
+      "cov@0.5": 0.9200603318250377,
+      "acc@0.5": 0.8326844262295082,
+      "cov@0.7": 0.7074849170437406,
+      "acc@0.7": 0.9077948034643571,
+      "cov@0.9": 0.4572963800904977,
+      "acc@0.9": 0.9647495361781077
+    },
+    "after": {
+      "n": 10608,
+      "acc": 0.7931749622926093,
+      "nll": 0.5079944898473756,
+      "brier": 0.2647027577519989,
+      "ece": 0.009998726121221617,
+      "mean_conf": 0.7984406564077068,
+      "score_mae": 0.3859622411504242,
+      "cov@0.5": 0.9168552036199095,
+      "acc@0.5": 0.8339502364795394,
+      "cov@0.7": 0.7034313725490197,
+      "acc@0.7": 0.908201554543018,
+      "cov@0.9": 0.4518288084464555,
+      "acc@0.9": 0.9670352597538077
+    },
+    "tasks": [
+      "ag_news",
+      "banking77",
+      "boolq",
+      "clinc150",
+      "commonsense_qa",
+      "copa",
+      "dbpedia",
+      "emotion",
+      "facts",
+      "fits",
+      "imdb",
+      "jailbreak",
+      "massive_intent",
+      "mnli",
+      "mrpc",
+      "openbookqa",
+      "paws",
+      "qnli",
+      "read",
+      "rte",
+      "sciq",
+      "sst2",
+      "swag",
+      "tweet_emoji",
+      "tweet_hate",
+      "tweet_irony",
+      "tweet_offensive",
+      "tweet_sentiment",
+      "yahoo",
+      "yelp",
+      "anli",
+      "winogrande",
+      "hellaswag",
+      "race",
+      "scitail",
+      "qqp",
+      "stsb",
+      "toxic",
+      "stance_abortion",
+      "stance_atheism",
+      "stance_feminist",
+      "stance_hillary",
+      "match",
+      "goemo_soft",
+      "reason",
+      "formality",
+      "politeness",
+      "strategyqa",
+      "vitaminc",
+      "ruletaker",
+      "proofwriter",
+      "folio",
+      "logiqa",
+      "tracie",
+      "temporal_nli",
+      "piqa",
+      "siqa",
+      "clutrr",
+      "gsm8k",
+      "svamp",
+      "aqua"
+    ]
+  }
+}
\ No newline at end of file