Add audited original IR control dataset

Adds ir-control-v1 with 11,600 original typed records, 8,800 request cases, 400 queries and 200 system families in five splits. Includes exact raw JSONL, Parquet, original manifest, queries/cases, independent audit and pinned source links. No IR training or TREC evaluation is claimed; a separate six-query Jev pilot is documented. Existing payloads, configs, manifests, default and frozen training mixtures are preserved.

README.md

@@ -135,11 +135,23 @@ configs:
         path: data/silent-failure-control-v1/test-*.parquet
       - split: ood
         path: data/silent-failure-control-v1/ood-*.parquet
+  - config_name: ir-control-v1
+    data_files:
+      - split: train
+        path: data/ir-control-v1/train-*.parquet
+      - split: calibration
+        path: data/ir-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/ir-control-v1/validation-*.parquet
+      - split: test
+        path: data/ir-control-v1/test-*.parquet
+      - split: ood
+        path: data/ir-control-v1/ood-*.parquet
 ---
 
 # Open-Jev: typed decision datasets
 
-Open-Jev turns a state and a question into a typed decision: a yes/no probability, a distribution over choices, independent label probabilities, or a discrete numeric/ordinal decision. This repository publishes ten separate, frozen data configs from the [Open-Jev project](https://github.com/Zefan-Cai/Open-Jev-Dev), together with original manifests, exact raw records, source code and reconstruction instructions.
+Open-Jev turns a state and a question into a typed decision: a yes/no probability, a distribution over choices, independent label probabilities, or a discrete numeric/ordinal decision. This repository publishes eleven separate, frozen data configs from the [Open-Jev project](https://github.com/Zefan-Cai/Open-Jev-Dev), together with original manifests, exact raw records, source code and reconstruction instructions.
 
 These are controlled, mostly synthetic tasks and reference labels. They are not official TypeSafe/Jev training data, model predictions, or evidence of general capability. Open-Jev is independently implemented and is not affiliated with TypeSafe.
 
@@ -157,6 +169,7 @@ These are controlled, mostly synthetic tasks and reference labels. They are not
 | `context-retention-control-v1` | 6,138 | 456 | 558 | 522 | 2,160 | 9,834 |
 | `sponsor-segment-control-v1` | 6,345 | 756 | 540 | 999 | 2,160 | 10,800 |
 | `silent-failure-control-v1` | 6,432 | 264 | 360 | 624 | 1,920 | 9,600 |
+| `ir-control-v1` | 7,366 | 464 | 464 | 986 | 2,320 | 11,600 |
 
 **The configs overlap.** `release-v2-redistributable` is contained in `browser-drone-expansion-v1-redistributable`; do not add config totals and call them unique examples. Counts are typed decision rows. Several heads may come from the same conversation, document, family or game trajectory.
 
@@ -190,7 +203,7 @@ For reproducibility, pass `revision="<dataset commit SHA>"`. Choose a config exp
 | `record_json` | Complete original JSON record, retaining original object key order and numeric representation. |
 | `original_line_number` | One-based row position in the original frozen split, including positions of excluded rows. |
 
-The Parquet representation avoids imposing one nested schema on different task states. `raw/<config>/<split>.jsonl.gz` preserves the source JSONL bytes after decompression. For the filtered composites, retained lines preserve exact bytes and order. For the eight stand-alone control configs, decompressed files match the original frozen split hashes exactly.
+The Parquet representation avoids imposing one nested schema on different task states. `raw/<config>/<split>.jsonl.gz` preserves the source JSONL bytes after decompression. For the filtered composites, retained lines preserve exact bytes and order. For the nine stand-alone control configs, decompressed files match the original frozen split hashes exactly.
 
 Use only `state`, `question`, `kind` and `options` as model inputs. Do not expose `target`, `metadata`, identities, split assignments or provenance fields to the model. Distribution, binary, multilabel and ordinal targets have different semantics; do not reduce every row to a single-class accuracy calculation.
 
@@ -227,7 +240,7 @@ Wikispeedia references:
 
 ## Community task additions — 2026-09-20
 
-Three additional original CC0 control configs contribute **30,234 typed rows across 15 new splits**. The repository now declares **10 configs and 50 splits**. Existing configs, payloads, manifests and the default remain unchanged.
+Three additional original CC0 control configs contribute **30,234 typed rows across 15 new splits**. That addition brought the repository to **10 configs and 50 splits**. Existing configs, payloads, manifests and the default remain unchanged.
 
 **These three additions have not been used for training or model evaluation.** Independent data audits validate their reference labels and split integrity; they are not model benchmarks. The released 2B/9B training mixture is unchanged.
 
@@ -236,3 +249,23 @@ Three additional original CC0 control configs contribute **30,234 typed rows acr
 - [`silent-failure-control-v1`](cards/silent-failure-control-v1.md): Ask one is_silent_failure Noul using only the exact response body string. The transport status is outside model input. Current maintenance, business rejection, a sign-in page replacing requested data, missing explicitly required receipts and unmet delivery requirements are contrasted with valid empty results, recovered history, quoted error text, accepted queued jobs and permitted partial results. Judge transport failures never become negative ground truth. Whole provider/contract families retain all counterfactuals; OOD reserves Chinese wording and different JSON/HTML/text layouts. This controlled grammar does not establish correctness for arbitrary APIs.
 
 The new original cases remain separate audit artifacts. Exact raw JSONL bytes, original manifests and an isolated source snapshot accompany the Parquet splits. [The additive export manifest](exports/community-data-20260920.json) records only these additions; the original `export-manifest.json` is retained byte for byte.
+
+
+## Original retrieval-control addition — 2026-09-20
+
+[`ir-control-v1`](cards/ir-control-v1.md) adds **11,600 typed rows across five
+splits**, representing 8,800 requests, 400 queries and 200 complete fictional
+system families. The repository now declares **11 configs and 55 splits**.
+It supports pointwise Noul/Score, pairwise/setwise Choice and listwise
+Choice/Score task forms under an original four-level relevance rubric.
+Choice labels supervise best-passage selection, not full ranking or uncertainty.
+
+The full corpus passed an independent body-derived label audit. No training on
+these IR records is claimed. A separately frozen six-query Jev pilot is documented
+in the source project; it is not a full-corpus or TREC evaluation. No TREC data,
+qrels, provider responses or third-party examples are included. Frozen pilot
+test/OOD families remain held out in the expanded corpus.
+
+All existing dataset payloads, original manifests, config definitions, the default
+config and frozen model-training mixtures remain unchanged. The
+[IR export manifest](exports/ir-data-20260920.json) binds this addition only.

artifacts/ir-control-v1/cases.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd667492b14bc813a74093cf1cc8529e51a9ad78df5aaf46472463b9af2b7c16
+size 302845

artifacts/ir-control-v1/queries.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc182ee526e7c09b55531dc36ad7a1c981e1af6b20b2cb5aaaa9f1c7c758e230
+size 55182

cards/ir-control-v1.md

@@ -0,0 +1,68 @@
+# Original graded retrieval controls
+
+**11,600 typed rows, 8,800 request cases, 400 queries and 200 synthetic system families.**
+Original generated content and labels are CC0-1.0. Source code is MIT.
+This is a finite original relevance task, not TREC or scraped search data.
+
+Each query asks for the current retry ceiling and backoff delay for one fictional
+system/profile. Grade 3 supplies both values; grade 2 supplies one; grade 1 is
+about the requested system but has the wrong profile, withdrawn guidance or no
+requested value; grade 0 concerns another system. A quoted query does not change
+the passage's subject. Two counterfactual queries share each family's eight passages.
+
+Six request forms are supplied: pointwise Noul/Score, opposite-order pairwise
+Choice, local setwise Choice, listwise Choice and listwise Score. Each query has
+22 request cases and 29 typed rows. One listwise Score request has eight questions.
+Score targets are four-level hard grades. Noul is positive only for a complete
+grade-3 answer. Choice targets are uniform over equally best passages; they do
+not supervise a full ranking or represent measured model uncertainty.
+
+| Split | Typed rows |
+| --- | ---: |
+| train | 7,366 |
+| calibration | 464 |
+| validation | 464 |
+| test | 986 |
+| ood | 2,320 |
+
+All query variants, passages and methods from a system family stay together.
+OOD reserves brief wording and different profile names for 40 whole families.
+Family index modulo 10 values 8–9 designate OOD; other families use stable group
+hashing. Expansion preserves every frozen pilot family assignment. No pilot
+test/OOD family enters train, calibration or validation. Rows within a family
+are correlated and must not be presented as independent real-world examples.
+
+Use only state, question, kind and options as model inputs. Targets, metadata,
+queries' reference relevance and auxiliary cases are privileged audit data.
+Each Parquet `record_json` and the decompressed raw JSONL preserve the original
+row bytes and order exactly, using the root card's existing column schema.
+
+**No training on this corpus is claimed.** A separate frozen pilot received
+real Jev 1.13.0 outputs on six held-out queries (132 requests / 174 typed
+answers). That small pilot is described in the linked source documentation;
+it is not evaluation of all 11,600 rows, a TREC reproduction, or evidence of
+newly trained Open-Jev capability. Generation manifests' inference flags record
+generation-time operations, not the absence of subsequent pilot evaluation.
+The frozen release-v2 mixture and published model training data remain unchanged.
+
+The [community interface source](https://github.com/ielab/llm-rankers/blob/ac843ed63a302900d76722b83be926fdb84a3936/jev/jev_rankers.py)
+supplies task shapes. Our prompts, passages and labels are independently authored.
+No third-party examples, TREC passages/qrels or provider responses are included.
+
+- [Original manifest](../provenance/original-manifests/ir-control-v1.json)
+- [Independent body audit](../provenance/ir-data-20260920/independent-audit.json)
+- [Frozen-pilot split preservation](../provenance/ir-data-20260920/stable-split-verification.json)
+- [Auxiliary queries](../artifacts/ir-control-v1/queries.jsonl.gz) and [request cases](../artifacts/ir-control-v1/cases.jsonl.gz)
+- [Pinned generator](https://github.com/Zefan-Cai/Open-Jev/blob/887b9fe3df729310014ac479c8eb7aad4d52bfbd/jev/ir_data.py) and [data/runtime documentation](https://github.com/Zefan-Cai/Open-Jev/blob/887b9fe3df729310014ac479c8eb7aad4d52bfbd/docs/ir-control-data.md)
+- [Incremental export manifest](../exports/ir-data-20260920.json)
+
+From that pinned source checkout, use a fresh output directory:
+
+```bash
+python3 -m jev.ir_data --output-dir data/ir-control-v1-rebuilt --groups 200 --ood-groups 40 --seed 42
+python3 reports/ir-control-v1/verify.py --data data/ir-control-v1-rebuilt --output ir-audit.json
+```
+
+External TREC evaluation remains pending. The local holdout loader isolates
+external candidates/qrels from the training-data tree and uses linear-gain
+`trec_eval ndcg_cut` semantics; no external benchmark result is bundled here.

data/ir-control-v1/calibration-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:864be79af98ec1145637b42882427f9f31e1de55628255b58eb19e6809563217
+size 46180

data/ir-control-v1/ood-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52062c0f76afe608df91d780be739c67330cd7f4b538470216f0b3fe662fe1da
+size 232768

data/ir-control-v1/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:082495de2ca3c8487349c06df8f3bc6bcf6ff8128ec729616fd1eed146ee2c56
+size 92585

data/ir-control-v1/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0788eba891c706dc2fd8f92d95f7172d0f5a482ab7d899d4cde312d23b9b9a5
+size 691414

data/ir-control-v1/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b95b2421e6cb6db45cac0bff10a58e8770789e6a973cf280fde99d25a81f7f7
+size 45870

exports/ir-data-20260920.json

@@ -0,0 +1,210 @@
+{
+  "schema_version": 1,
+  "repo_id": "ZefanCai/Open-Jev",
+  "release": "ir-data-20260920",
+  "parent_dataset_revision": "cad3d65e934ca0f1aede72b5c39d3f0384c4ca7f",
+  "operation": "Add ir-control-v1 five splits and its exact raw/auxiliary records, manifests, audit and card; update root card only",
+  "new_data_training_performed": false,
+  "full_corpus_model_evaluation_performed": false,
+  "separate_frozen_pilot_evaluation": {
+    "provider": "Jev 1.13.0",
+    "queries": 6,
+    "groups": 3,
+    "requests": 132,
+    "typed_answers": 174,
+    "raw_model_responses_included": false
+  },
+  "external_trec_downloaded": false,
+  "external_trec_evaluated": false,
+  "frozen_training_mixture_changed": false,
+  "frozen_pilot_manifest_sha256": "beb827c63486e776d6d1abc0d178f557394dbebe57de72523ff4c3112f17d73b",
+  "source_commit": "887b9fe3df729310014ac479c8eb7aad4d52bfbd",
+  "source_blob_base_url": "https://github.com/Zefan-Cai/Open-Jev/blob/887b9fe3df729310014ac479c8eb7aad4d52bfbd",
+  "source_generator_sha256": "c615f711e57ccf5488468de3aa649447c2bc8d4c884e4aa42c820459b9b781d0",
+  "configs": {
+    "ir-control-v1": {
+      "total": 11600,
+      "groups": 200,
+      "query_count": 400,
+      "request_case_count": 8800,
+      "license": "CC0-1.0",
+      "original_manifest_sha256": "5305dad7326fbf32aa006447137c5a3eb92872d90cde71302bc7efb39f882590",
+      "splits": {
+        "train": {
+          "count": 7366,
+          "original_sha256": "6d33876576fb7989abbf9edcd878074f4da7d4f502d770d110a848bbe259541d",
+          "raw_path": "raw/ir-control-v1/train.jsonl.gz",
+          "raw_uncompressed_sha256": "6d33876576fb7989abbf9edcd878074f4da7d4f502d770d110a848bbe259541d",
+          "raw_sha256": "b55297bd98bb187fb5fa71df92382ffb23fecbbd9234f1c185491b0eb4ff9dd3",
+          "parquet_path": "data/ir-control-v1/train-00000-of-00001.parquet",
+          "parquet_sha256": "e0788eba891c706dc2fd8f92d95f7172d0f5a482ab7d899d4cde312d23b9b9a5",
+          "original_line_numbers_preserved": true,
+          "sources": {
+            "ir-control-v1": 7366
+          },
+          "kinds": {
+            "noul": 2032,
+            "score": 4064,
+            "choice": 1270
+          }
+        },
+        "calibration": {
+          "count": 464,
+          "original_sha256": "7ad15592ec8b1d2e6c2d819fb186dc2dec47f8ed9214072f6f45452be7194e61",
+          "raw_path": "raw/ir-control-v1/calibration.jsonl.gz",
+          "raw_uncompressed_sha256": "7ad15592ec8b1d2e6c2d819fb186dc2dec47f8ed9214072f6f45452be7194e61",
+          "raw_sha256": "14ad79c347cbfd1d036b10e3b90ecd1cf3ff6af58c42ba525a5ec4560a2c505e",
+          "parquet_path": "data/ir-control-v1/calibration-00000-of-00001.parquet",
+          "parquet_sha256": "864be79af98ec1145637b42882427f9f31e1de55628255b58eb19e6809563217",
+          "original_line_numbers_preserved": true,
+          "sources": {
+            "ir-control-v1": 464
+          },
+          "kinds": {
+            "noul": 128,
+            "score": 256,
+            "choice": 80
+          }
+        },
+        "validation": {
+          "count": 464,
+          "original_sha256": "0ecbe867ae894a83101805328be1b1d983cc210fb624fa78d2aedd59f8b702ec",
+          "raw_path": "raw/ir-control-v1/validation.jsonl.gz",
+          "raw_uncompressed_sha256": "0ecbe867ae894a83101805328be1b1d983cc210fb624fa78d2aedd59f8b702ec",
+          "raw_sha256": "2018ee228a230589061a43af8151531988d2f7d866ab4326048eaad8ed18c2f6",
+          "parquet_path": "data/ir-control-v1/validation-00000-of-00001.parquet",
+          "parquet_sha256": "4b95b2421e6cb6db45cac0bff10a58e8770789e6a973cf280fde99d25a81f7f7",
+          "original_line_numbers_preserved": true,
+          "sources": {
+            "ir-control-v1": 464
+          },
+          "kinds": {
+            "noul": 128,
+            "score": 256,
+            "choice": 80
+          }
+        },
+        "test": {
+          "count": 986,
+          "original_sha256": "97637e12062c98447b49cfa8cc9669e86525dce0078657a08fd9f1b54e5e2869",
+          "raw_path": "raw/ir-control-v1/test.jsonl.gz",
+          "raw_uncompressed_sha256": "97637e12062c98447b49cfa8cc9669e86525dce0078657a08fd9f1b54e5e2869",
+          "raw_sha256": "2db7cf46a70808e48f1f06adfdcc295b825644ebefcca7ee7838abfa3e5461de",
+          "parquet_path": "data/ir-control-v1/test-00000-of-00001.parquet",
+          "parquet_sha256": "082495de2ca3c8487349c06df8f3bc6bcf6ff8128ec729616fd1eed146ee2c56",
+          "original_line_numbers_preserved": true,
+          "sources": {
+            "ir-control-v1": 986
+          },
+          "kinds": {
+            "noul": 272,
+            "score": 544,
+            "choice": 170
+          }
+        },
+        "ood": {
+          "count": 2320,
+          "original_sha256": "fa6332b13395ef37ab4c88e4c5888201f8c814fdd26d387bc5684e7bf410076b",
+          "raw_path": "raw/ir-control-v1/ood.jsonl.gz",
+          "raw_uncompressed_sha256": "fa6332b13395ef37ab4c88e4c5888201f8c814fdd26d387bc5684e7bf410076b",
+          "raw_sha256": "bfa8061fd82e67da8d1a6facff72e2e0f017cc3c74351e6362c4cbd04b2555b8",
+          "parquet_path": "data/ir-control-v1/ood-00000-of-00001.parquet",
+          "parquet_sha256": "52062c0f76afe608df91d780be739c67330cd7f4b538470216f0b3fe662fe1da",
+          "original_line_numbers_preserved": true,
+          "sources": {
+            "ir-control-v1": 2320
+          },
+          "kinds": {
+            "noul": 640,
+            "score": 1280,
+            "choice": 400
+          }
+        }
+      },
+      "extra_artifacts": {
+        "artifacts/ir-control-v1/queries.jsonl.gz": {
+          "uncompressed_sha256": "7b5531d2efc5a650cf240fc57156774dc556b23d300d1d48a5302d7b046ee106",
+          "uncompressed_bytes": 1142734,
+          "sha256": "bc182ee526e7c09b55531dc36ad7a1c981e1af6b20b2cb5aaaa9f1c7c758e230",
+          "bytes": 55182
+        },
+        "artifacts/ir-control-v1/cases.jsonl.gz": {
+          "uncompressed_sha256": "2261aea158332b8831a5bd8547bc39cb178671ebc5685a4ecfe8360fdaa6f492",
+          "uncompressed_bytes": 14235110,
+          "sha256": "dd667492b14bc813a74093cf1cc8529e51a9ad78df5aaf46472463b9af2b7c16",
+          "bytes": 302845
+        }
+      }
+    }
+  },
+  "published_files": {
+    "README.md": {
+      "sha256": "12670b1116e6b42bfae177d6be91a0aea1c1deff1f093816ae2b6f640ab68188",
+      "bytes": 19003
+    },
+    "artifacts/ir-control-v1/cases.jsonl.gz": {
+      "sha256": "dd667492b14bc813a74093cf1cc8529e51a9ad78df5aaf46472463b9af2b7c16",
+      "bytes": 302845
+    },
+    "artifacts/ir-control-v1/queries.jsonl.gz": {
+      "sha256": "bc182ee526e7c09b55531dc36ad7a1c981e1af6b20b2cb5aaaa9f1c7c758e230",
+      "bytes": 55182
+    },
+    "cards/ir-control-v1.md": {
+      "sha256": "57518acd272f546403f0c6b05d34ff789b7bf9f38c978506d813d29d8c3fb717",
+      "bytes": 4133
+    },
+    "data/ir-control-v1/calibration-00000-of-00001.parquet": {
+      "sha256": "864be79af98ec1145637b42882427f9f31e1de55628255b58eb19e6809563217",
+      "bytes": 46180
+    },
+    "data/ir-control-v1/ood-00000-of-00001.parquet": {
+      "sha256": "52062c0f76afe608df91d780be739c67330cd7f4b538470216f0b3fe662fe1da",
+      "bytes": 232768
+    },
+    "data/ir-control-v1/test-00000-of-00001.parquet": {
+      "sha256": "082495de2ca3c8487349c06df8f3bc6bcf6ff8128ec729616fd1eed146ee2c56",
+      "bytes": 92585
+    },
+    "data/ir-control-v1/train-00000-of-00001.parquet": {
+      "sha256": "e0788eba891c706dc2fd8f92d95f7172d0f5a482ab7d899d4cde312d23b9b9a5",
+      "bytes": 691414
+    },
+    "data/ir-control-v1/validation-00000-of-00001.parquet": {
+      "sha256": "4b95b2421e6cb6db45cac0bff10a58e8770789e6a973cf280fde99d25a81f7f7",
+      "bytes": 45870
+    },
+    "provenance/ir-data-20260920/independent-audit.json": {
+      "sha256": "9fcc50de615a74d6f5b25444a0b77d4c4cfc09bd4bde1ad055c5c0624ebd948a",
+      "bytes": 1172
+    },
+    "provenance/ir-data-20260920/stable-split-verification.json": {
+      "sha256": "0dcc91e9038dea768c020cb9f5c8617892cca04c5a935de564cf28ca7b6807e8",
+      "bytes": 3208
+    },
+    "provenance/original-manifests/ir-control-v1.json": {
+      "sha256": "5305dad7326fbf32aa006447137c5a3eb92872d90cde71302bc7efb39f882590",
+      "bytes": 2577
+    },
+    "raw/ir-control-v1/calibration.jsonl.gz": {
+      "sha256": "14ad79c347cbfd1d036b10e3b90ecd1cf3ff6af58c42ba525a5ec4560a2c505e",
+      "bytes": 17586
+    },
+    "raw/ir-control-v1/ood.jsonl.gz": {
+      "sha256": "bfa8061fd82e67da8d1a6facff72e2e0f017cc3c74351e6362c4cbd04b2555b8",
+      "bytes": 86800
+    },
+    "raw/ir-control-v1/test.jsonl.gz": {
+      "sha256": "2db7cf46a70808e48f1f06adfdcc295b825644ebefcca7ee7838abfa3e5461de",
+      "bytes": 36298
+    },
+    "raw/ir-control-v1/train.jsonl.gz": {
+      "sha256": "b55297bd98bb187fb5fa71df92382ffb23fecbbd9234f1c185491b0eb4ff9dd3",
+      "bytes": 265316
+    },
+    "raw/ir-control-v1/validation.jsonl.gz": {
+      "sha256": "2018ee228a230589061a43af8151531988d2f7d866ab4326048eaad8ed18c2f6",
+      "bytes": 17584
+    }
+  }
+}

provenance/ir-data-20260920/independent-audit.json

@@ -0,0 +1,50 @@
+{
+  "status": "passed",
+  "summary": {
+    "records": 11600,
+    "groups": 200,
+    "unique_inputs": 11600,
+    "splits": {
+      "calibration": 464,
+      "ood": 2320,
+      "test": 986,
+      "train": 7366,
+      "validation": 464
+    },
+    "kinds": {
+      "choice": 2000,
+      "noul": 3200,
+      "score": 6400
+    },
+    "families": {
+      "evidence": 11600
+    },
+    "sources": {
+      "ir-control-v1": 11600
+    }
+  },
+  "queries": 400,
+  "request_cases": 8800,
+  "families": 200,
+  "request_methods": {
+    "pointwise_noul": 3200,
+    "pointwise_score": 3200,
+    "pairwise": 800,
+    "setwise": 800,
+    "listwise_choice": 400,
+    "listwise_score": 400
+  },
+  "body_derived_relevance_counts": {
+    "3": 400,
+    "0": 800,
+    "1": 1600,
+    "2": 400
+  },
+  "manifest_sha256": "5305dad7326fbf32aa006447137c5a3eb92872d90cde71302bc7efb39f882590",
+  "generator_imported": false,
+  "reference_metadata_used_to_derive_labels": false,
+  "external_trec_downloaded": false,
+  "external_trec_evaluated": false,
+  "model_inference_performed": false,
+  "scope": "Original finite relevance grammar only; not real-world relevance judgments or reproduced TREC metrics."
+}

provenance/ir-data-20260920/stable-split-verification.json

@@ -0,0 +1,119 @@
+{
+  "status": "passed",
+  "pilot_manifest_sha256": "beb827c63486e776d6d1abc0d178f557394dbebe57de72523ff4c3112f17d73b",
+  "full_manifest_sha256": "5305dad7326fbf32aa006447137c5a3eb92872d90cde71302bc7efb39f882590",
+  "all_seven_pilot_jsonl_byte_identical_after_stable_split_fix": {
+    "train.jsonl": true,
+    "calibration.jsonl": true,
+    "validation.jsonl": true,
+    "test.jsonl": true,
+    "ood.jsonl": true,
+    "queries.jsonl": true,
+    "cases.jsonl": true
+  },
+  "pilot_family_assignments_preserved": 10,
+  "protected_pilot_test_ood_families": {
+    "ir:b9bf64bec231": "test",
+    "ir:d717b19ba847": "ood",
+    "ir:b0db813958f1": "ood"
+  },
+  "protected_families_in_full_training_calibration_validation": 0,
+  "pilot_audit": {
+    "status": "passed",
+    "summary": {
+      "records": 580,
+      "groups": 10,
+      "unique_inputs": 580,
+      "splits": {
+        "ood": 116,
+        "test": 58,
+        "train": 406
+      },
+      "kinds": {
+        "choice": 100,
+        "noul": 160,
+        "score": 320
+      },
+      "families": {
+        "evidence": 580
+      },
+      "sources": {
+        "ir-control-v1": 580
+      }
+    },
+    "queries": 20,
+    "request_cases": 440,
+    "families": 10,
+    "request_methods": {
+      "pointwise_noul": 160,
+      "pointwise_score": 160,
+      "pairwise": 40,
+      "setwise": 40,
+      "listwise_choice": 20,
+      "listwise_score": 20
+    },
+    "body_derived_relevance_counts": {
+      "3": 20,
+      "0": 40,
+      "1": 80,
+      "2": 20
+    },
+    "manifest_sha256": "beb827c63486e776d6d1abc0d178f557394dbebe57de72523ff4c3112f17d73b",
+    "generator_imported": false,
+    "reference_metadata_used_to_derive_labels": false,
+    "external_trec_downloaded": false,
+    "external_trec_evaluated": false,
+    "model_inference_performed": false,
+    "scope": "Original finite relevance grammar only; not real-world relevance judgments or reproduced TREC metrics."
+  },
+  "full_audit": {
+    "status": "passed",
+    "summary": {
+      "records": 11600,
+      "groups": 200,
+      "unique_inputs": 11600,
+      "splits": {
+        "calibration": 464,
+        "ood": 2320,
+        "test": 986,
+        "train": 7366,
+        "validation": 464
+      },
+      "kinds": {
+        "choice": 2000,
+        "noul": 3200,
+        "score": 6400
+      },
+      "families": {
+        "evidence": 11600
+      },
+      "sources": {
+        "ir-control-v1": 11600
+      }
+    },
+    "queries": 400,
+    "request_cases": 8800,
+    "families": 200,
+    "request_methods": {
+      "pointwise_noul": 3200,
+      "pointwise_score": 3200,
+      "pairwise": 800,
+      "setwise": 800,
+      "listwise_choice": 400,
+      "listwise_score": 400
+    },
+    "body_derived_relevance_counts": {
+      "3": 400,
+      "0": 800,
+      "1": 1600,
+      "2": 400
+    },
+    "manifest_sha256": "5305dad7326fbf32aa006447137c5a3eb92872d90cde71302bc7efb39f882590",
+    "generator_imported": false,
+    "reference_metadata_used_to_derive_labels": false,
+    "external_trec_downloaded": false,
+    "external_trec_evaluated": false,
+    "model_inference_performed": false,
+    "scope": "Original finite relevance grammar only; not real-world relevance judgments or reproduced TREC metrics."
+  }
+}

provenance/original-manifests/ir-control-v1.json

@@ -0,0 +1,76 @@
+{
+  "schema_version": 1,
+  "configuration": {
+    "type": "synthetic",
+    "generator_version": "ir-control-v1",
+    "groups": 200,
+    "ood_groups": 40,
+    "ood_assignment": "index_mod_10_greater_or_equal_8",
+    "seed": 42,
+    "license": "CC0-1.0",
+    "split_policy": "whole_system_family_all_query_mode_counterfactuals_and_passages; reserved_brief_wording_ood",
+    "source_interface": "https://github.com/ielab/llm-rankers/blob/ac843ed63a302900d76722b83be926fdb84a3936/jev/jev_rankers.py",
+    "source_code_or_examples_imported": false,
+    "generator_sha256": "c615f711e57ccf5488468de3aa649447c2bc8d4c884e4aa42c820459b9b781d0"
+  },
+  "summary": {
+    "records": 11600,
+    "groups": 200,
+    "unique_inputs": 11600,
+    "splits": {
+      "calibration": 464,
+      "ood": 2320,
+      "test": 986,
+      "train": 7366,
+      "validation": 464
+    },
+    "kinds": {
+      "choice": 2000,
+      "noul": 3200,
+      "score": 6400
+    },
+    "families": {
+      "evidence": 11600
+    },
+    "sources": {
+      "ir-control-v1": 11600
+    }
+  },
+  "files_sha256": {
+    "train.jsonl": "6d33876576fb7989abbf9edcd878074f4da7d4f502d770d110a848bbe259541d",
+    "calibration.jsonl": "7ad15592ec8b1d2e6c2d819fb186dc2dec47f8ed9214072f6f45452be7194e61",
+    "validation.jsonl": "0ecbe867ae894a83101805328be1b1d983cc210fb624fa78d2aedd59f8b702ec",
+    "test.jsonl": "97637e12062c98447b49cfa8cc9669e86525dce0078657a08fd9f1b54e5e2869",
+    "ood.jsonl": "fa6332b13395ef37ab4c88e4c5888201f8c814fdd26d387bc5684e7bf410076b",
+    "queries.jsonl": "7b5531d2efc5a650cf240fc57156774dc556b23d300d1d48a5302d7b046ee106",
+    "cases.jsonl": "2261aea158332b8831a5bd8547bc39cb178671ebc5685a4ecfe8360fdaa6f492"
+  },
+  "model_input_fields": [
+    "state",
+    "question",
+    "kind",
+    "options"
+  ],
+  "query_count": 400,
+  "request_case_count": 8800,
+  "request_methods": {
+    "pointwise_noul": 3200,
+    "pointwise_score": 3200,
+    "pairwise": 800,
+    "setwise": 800,
+    "listwise_choice": 400,
+    "listwise_score": 400
+  },
+  "relevance_counts": {
+    "3": 400,
+    "1": 1600,
+    "2": 400,
+    "0": 800
+  },
+  "training_performed": false,
+  "model_inference_performed": false,
+  "external_trec_downloaded": false,
+  "external_trec_evaluated": false,
+  "scope": "Original finite graded-relevance controls; query families and counterfactuals are correlated. Not a TREC reproduction.",
+  "target_semantics": "Score: four-level hard grade. Noul: both requested current values are supplied. Choice: uniform over equally best passages; not measured model uncertainty or full ranking supervision."
+}

raw/ir-control-v1/calibration.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14ad79c347cbfd1d036b10e3b90ecd1cf3ff6af58c42ba525a5ec4560a2c505e
+size 17586

raw/ir-control-v1/ood.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bfa8061fd82e67da8d1a6facff72e2e0f017cc3c74351e6362c4cbd04b2555b8
+size 86800

raw/ir-control-v1/test.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2db7cf46a70808e48f1f06adfdcc295b825644ebefcca7ee7838abfa3e5461de
+size 36298

raw/ir-control-v1/train.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b55297bd98bb187fb5fa71df92382ffb23fecbbd9234f1c185491b0eb4ff9dd3
+size 265316

raw/ir-control-v1/validation.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2018ee228a230589061a43af8151531988d2f7d866ab4326048eaad8ed18c2f6
+size 17584