Release audited Open-Jev 27B v1.1 adapter and decision head

LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 Alibaba Cloud
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

LICENSE-CODE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Open-Jev contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -0,0 +1,153 @@
+---
+license: apache-2.0
+base_model: Qwen/Qwen3.8-27B
+base_model_relation: adapter
+library_name: peft
+datasets:
+- ZefanCai/Open-Jev-v1.1
+tags:
+- open-jev
+- qwen3.8
+- lora
+- non-generative
+- typed-decisions
+---
+
+# Open-Jev-27B-v1.1
+
+**A trained LoRA adapter and scalar decision head for the text backbone of Qwen/Qwen3.8-27B.** The package requires the exact upstream model and tokenizer revision `1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0` and the [Open-Jev loader](https://github.com/Zefan-Cai/Open-Jev). It contains 15,471,617 trained parameters; upstream base-model weights are obtained separately.
+
+**Evaluation complete:** all 127,787 original-mixture Test/OOD rows were evaluated and independently audited, with no missing, duplicate or failed predictions. The same checkpoint scored **197/231 (85.28%)** on the public JevBench subset and **80/111 (72.07%)** on its Hard tier. Full results, exact denominators and separate evaluation protocols are below.
+
+Open-Jev scores caller-supplied candidates directly and returns typed decisions:
+
+- **Choice:** a probability distribution over the supplied candidates.
+- **Noul:** a yes/no probability.
+- **Score:** probabilities over supplied ordinal levels and their expected value.
+
+One LoRA adapter and scalar head serve these tasks. Each candidate receives a scalar score, so the number and meaning of candidates can change between requests and domains. The probabilities come from the decision head and saved temperature, without autoregressive answer generation. The server returns decisions; it does not execute actions.
+
+The package uses a text backbone. Gameplay and browser integrations supply structured observations or text through their task adapters. This release does not establish image-only perception or a full autonomous workflow for every domain.
+
+## Download and run
+
+Use an environment with sufficient GPU memory for the upstream model and selected input sizes. These commands use the conservative uncached path:
+
+```bash
+git clone https://github.com/Zefan-Cai/Open-Jev.git
+cd Open-Jev
+python -m pip install -e '.[train]'
+hf download ZefanCai/Open-Jev-27B-v1.1 --local-dir ./checkpoints/open-jev-27b-v1.1
+python -m jev.server \
+  --checkpoint ./checkpoints/open-jev-27b-v1.1/package/checkpoint \
+  --device cuda:0 --max-length 4096 --batch-size 1 --no-prefix-cache \
+  --host 127.0.0.1 --port 8791
+```
+
+The `train` extra also supplies inference dependencies: Transformers 5.10.2, PEFT 0.19.1, Accelerate 1.13.0, Torch and safetensors. Pin the downloaded model repository to a commit with `hf download --revision <commit>` for reproducible deployments.
+
+The Open-Jev loader reads the exact upstream revision from `package/checkpoint/model.json`. The unchanged PEFT adapter configuration has an empty `base_model_name_or_path`; a generic `AutoPeftModel` text-generation call is not the loading interface for this package. It would also omit the separately saved scalar head and temperature.
+
+Example request, provided as an input example rather than a recorded prediction:
+
+```bash
+curl http://127.0.0.1:8791/v1/systemone \
+  -H 'Content-Type: application/json' \
+  -d '{"state":"I was charged twice and want a refund.","questions":{"intent":{"type":"choice","instructions":"Choose the customer intent.","criteria":{"billing":"A payment or refund issue","technical":"A malfunction or setup issue","other":"Another request"}},"refund_requested":{"type":"noul","instructions":"Does the customer explicitly request a refund?"}}}'
+```
+
+The saved maximum input length is **4,096 tokens per candidate**. Overlength inputs are rejected rather than silently truncated. Larger experimental evaluation limits do not change the training length or establish equivalent quality at those lengths. Prefix-cache or cross-question batching experiments require their own checkpoint and hardware validation; this card makes no proprietary-kernel equivalence or GPU speedup claim.
+
+## Training
+
+This checkpoint completed a predeclared pass over the frozen `community-hard-mix-v2-final` training split:
+
+| Setting | Value |
+| --- | --- |
+| Unique training rows | 148,639 |
+| Optimizer steps for this mixture | 37,160 |
+| Global batch | 4 |
+| Consumed rows | 148,640; one row wraps at the final step |
+| Trainable parameters | 15,471,617 |
+| LoRA | Rank 8, alpha 16, dropout 0 |
+| LoRA targets | `q_proj`, `k_proj`, `v_proj`, `o_proj`, `in_proj_qkv`, `out_proj` |
+| LoRA / head learning rates | `2e-5` / `5e-5` |
+| Objective | Candidate negative log likelihood plus Brier loss, weight `0.1` |
+| Sampling / seed | Shuffled / `20260921` |
+| Maximum training input length | 4,096 tokens |
+| Saved temperature | `2.5343690298472983`, fitted on 512 calibration rows |
+| Final continuation code commit | `7ac6bab261bd97efc3d446152999e94ae432e5b0` |
+
+The run **warm-started from an earlier Open-Jev 27B checkpoint at step 26,500**, trained on the browser/drone expansion mixture. The new mixture started with fresh optimizer state, random streams and step cursor. Thus the 148,639 rows above describe this additional training stage, not all data ever used by the adapter. The source snapshot manifest hash is `28aca3c8977191f976b0b961bd9924e4ba4ea511c6b627a887ae7cad937f41a5`.
+
+The run later continued through a state-preserving migration to four-rank FSDP2 for the frozen BF16 backbone, with replicated FP32 LoRA/head parameters. This records execution provenance; it does not claim bitwise equality to uninterrupted DDP. The training-time baseline is the warm-start checkpoint, not raw Qwen.
+
+## Data and provenance
+
+The final training stage combines **82,045 WANLI NLI rows**, **30,720 new synthetic community-task rows**, and **35,874 replay rows**. The new community tasks cover support, browser tools, games, information retrieval, RAG, contracts, shell-history tasks and rubric judgments. Replay adds the earlier extraction, email, citation, workflow, navigation, painting and control tasks. The original training manifest names 74 source groups; these are not 74 independently validated real-world domains.
+
+The original frozen mixture contains 328,672 rows across all splits:
+
+| Split | Original mixture | Public redistributable projection |
+| --- | ---: | ---: |
+| Train | 148,639 | 147,139 |
+| Calibration | 26,764 | 26,675 |
+| Validation | 25,482 | 25,413 |
+| Test | 43,301 | 43,125 |
+| OOD | 84,486 | 84,267 |
+
+The [v1.1 dataset](https://huggingface.co/datasets/ZefanCai/Open-Jev-v1.1), configuration `community-hard-mix-v2-redistributable`, omits 2,053 original `wikispeedia-v1` rows, including 1,500 training rows, because redistribution permission for the source graph/path archive has not been confirmed. **The public projection is not byte-identical to the data used for these weights.** Internal evaluation of the original mixture includes those Wiki rows. The dataset's reconstruction metadata can restore the original mixture when the omitted files are separately supplied and verified.
+
+The original manifest SHA-256 is `7b26f948d2ae11f20d7a18be437d1ada616fc479e1596ae94be7596787fe4e54`; the original training-file SHA-256 is `959dac64adc1717d92e9bb80371d4b940b2b5994f9e36e469b3eaed85883fdff`. [The full manifest](provenance/original-training-mixture-manifest.json) and [provenance record](package/provenance.json) preserve split and component hashes. No dataset rows are bundled with this model.
+
+WANLI is credited to [alisawuffles/WANLI](https://github.com/alisawuffles/WANLI) at revision `61c95318fd71c55b6ba355d76253254615f387ec`, under CC-BY-4.0. Its source premise/hypothesis text and published labels were converted into the common decision format. New synthetic community data and the listed original synthetic control sources carry their own CC0-1.0 declarations. The browser/drone aggregate retains per-component terms, including the Wikispeedia limitation above. Dataset licenses do not relicense model weights.
+
+The saved lexical overlap screen found no matches against 531 frozen visible benchmark requests. This checks exact or lexical overlap only; it does not prove semantic independence, exclude paraphrases, or assess upstream pretraining exposure. Later community v3/v4 mixtures are not part of this completed training stage.
+
+## Evaluation results and limits
+
+The full internal evaluation covers **43,301 Test + 84,486 OOD = 127,787 rows** of the original frozen mixture. All rows have valid predictions; the independent audit found **zero missing, duplicate or failed records** and verified the exact checkpoint, dataset, code, saved-temperature probabilities, successful controller exits and cleanup. These are complete splits, without sampling.
+
+| Checkpoint | Old Test | Old OOD | Expanded Test | Expanded OOD |
+| --- | ---: | ---: | ---: | ---: |
+| Open-Jev-27B-v1.1 | **9,876 / 10,046 (98.31%)** | **14,825 / 15,446 (95.98%)** | **41,357 / 42,789 (96.65%)** | **80,934 / 83,924 (96.44%)** |
+
+Cells show correct hard labels / all hard labels. The complete raw panels contain 10,532 / 15,920 / 43,301 / 84,486 rows respectively. The expanded panels include **1,074 soft-label rows**; these remain in the probability metrics but are excluded from hard-label accuracy. Old Test/OOD are content-identical subsets of Expanded Test/OOD, so the panels overlap. Expanded scores cover the original complete mixture, including the Wiki rows omitted from the public data projection. The earlier 512-row training diagnostics are not used for any cell.
+
+The [complete internal audit report](verification/internal-full.json) includes hard/soft denominators, NLL, Brier, expected accuracy, ECE and evidence hashes. It was produced with a BF16 base, FP32 head, maximum length **4,096**, no truncation, no prefix cache, one row per rank, and the checkpoint's saved temperature **2.5343690298472983**; no temperature was fitted on the evaluation data. The report SHA-256 is `19a154c1a6ed3935bac266ab04b8a395bc180739540127b4b3802190e3c34f97`.
+
+### JevBench public subset
+
+The separately audited JevBench run uses **231 public tasks**: 72 original, 48 easy and 111 Hard. This is **not the complete 534-task benchmark**; unavailable private/judge tiers were not evaluated and no full-benchmark composite is reported.
+
+| Model | Public 231 | Hard 111 |
+| --- | ---: | ---: |
+| Released Open-Jev 2B | 150 / 231 (64.94%) | 46 / 111 (41.44%) |
+| Released Open-Jev 9B | 179 / 231 (77.49%) | 66 / 111 (59.46%) |
+| **Open-Jev-27B-v1.1** | **197 / 231 (85.28%)** | **80 / 111 (72.07%)** |
+| Jev 1.13.0 | 200 / 231 (86.58%) | 81 / 111 (72.97%) |
+
+27B v1.1 remains **three tasks behind Jev overall and one task behind on Hard**. This run used the identical adapter/head/checkpoint hash and saved temperature, with a **16,384-token** JevBench limit and no prefix cache or retries; it is a separate protocol from the 4,096-token internal evaluation. [The JevBench report](verification/jevbench.json) records task-family results, baseline counts and protocol details. New 2B training was stopped and new 9B training has not started; the smaller-model rows above refer to the previously released checkpoints.
+
+No latency or efficiency comparison is reported in this package. Shared-node four-rank collective execution timings and single-GPU HTTP or hosted API timings measure different execution conditions and cannot establish a direct speedup ratio.
+
+Typed-decision accuracy is not game win rate or end-to-end workflow completion. Synthetic holdouts, a single NLI corpus and lexical screening do not establish broad operational reliability or calibrated probabilities on arbitrary new tasks. The model is an independent Qwen-based Open-Jev implementation, not a release of Jev's private weights or sampler.
+
+## Artifact verification
+
+The package preserves the complete six-file checkpoint tree byte for byte, including the original generated adapter README:
+
+| Artifact | SHA-256 |
+| --- | --- |
+| LoRA adapter | `1c857224bd3609c6a71eacf7f71dd021115fcc0f791936b1fc332e915b548a81` |
+| Scalar head | `76e382f122abfa4e0c467d860a8d142d2fb6d2a98dc0ef9e19870bfc6eb296b4` |
+| Original evaluated checkpoint directory | `c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71` |
+| Packaged checkpoint directory | `c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71` |
+
+The source and packaged checkpoint directory hashes are identical, so the downloaded checkpoint retains the identity used by the fixed-checkpoint evaluator. The unchanged `adapter/README.md` is a generated PEFT template preserved for byte identity; this root card provides the model documentation. [Checkpoint verification](verification/checkpoint.json) records all six file hashes.
+
+[CPU validation](verification/cpu.json) checked 320 LoRA tensors plus the scalar head: all are finite FP32, and every adapter tensor shape matches the exact pinned Qwen text architecture instantiated on the meta device. It also verified calibration IDs and targets against the original calibration split. No base weights were materialized and no model inference was rerun during packaging. These checks establish structure and artifact identity, not a fresh 27B inference result.
+
+## License and attribution
+
+The trained adapter and head are distributed under **Apache-2.0** with the complete pinned Qwen/Alibaba Cloud license and attribution in [LICENSE](LICENSE) and [UPSTREAM.md](UPSTREAM.md). Open-Jev source code is **MIT**, preserved separately in [LICENSE-CODE](LICENSE-CODE). The upstream model/tokenizer must be obtained under their own terms. No endorsement by Alibaba Cloud, the Qwen Team or TypeSafe is implied.

UPSTREAM.md

@@ -0,0 +1,36 @@
+# Upstream attribution and modifications
+
+Copyright 2026 Alibaba Cloud.
+
+Open-Jev-27B-v1.1 uses the text backbone of **Qwen/Qwen3.8-27B**, pinned to
+`1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0`.
+
+- [Pinned upstream model card](https://huggingface.co/Qwen/Qwen3.8-27B/blob/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0/README.md)
+- [Pinned upstream license](https://huggingface.co/Qwen/Qwen3.8-27B/blob/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0/LICENSE)
+
+The included `LICENSE` preserves those exact Apache License 2.0 bytes and
+the Alibaba Cloud copyright notice. SHA-256:
+`bbedc3fda3305820b977265f01b8619d87570a6739de3a5582c3464840f1e57a`.
+The pinned model card and Hugging Face revision metadata separately identify
+Apache-2.0. The inspected upstream file inventory contains no `NOTICE` file.
+
+**Modifications by Open-Jev:** `package/checkpoint/adapter/` contains trained
+LoRA weights for the Qwen text backbone. `package/checkpoint/head.pt` is a
+trained scalar decision head whose lineage starts from the upstream
+Yes-minus-No output-embedding readout. `model.json` selects the pinned base
+and the Open-Jev decision interface; `temperature.json` stores calibration
+fitted on a separate split. These files are modified training products and
+are not unmodified Qwen weights. Their training-stage and initialization
+provenance is recorded in `package/provenance.json`.
+
+The inference weight package is Apache-2.0. Original Open-Jev source code is
+MIT; its separate license is included as `LICENSE-CODE`. Base weights,
+tokenizer files, third-party runtime libraries and training dataset rows
+are not bundled here.
+
+The pinned Qwen model card requests citation of: Qwen Team,
+*Qwen3.8-Max: A New Bar for Coding and Cowork*, August 2026,
+<https://qwen.ai/blog?id=qwen3.8>. This citation title comes from that model
+card and does not identify this 27B checkpoint as the hosted Qwen3.8-Max model.
+
+No endorsement by Alibaba Cloud, the Qwen Team or TypeSafe is implied.

package/checkpoint/adapter/README.md

@@ -0,0 +1,206 @@
+---
+base_model: ''
+library_name: peft
+tags:
+- 'base_model:adapter:'
+- 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 0.19.1

package/checkpoint/adapter/adapter_config.json

@@ -0,0 +1,50 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3_5TextModel",
+    "parent_library": "transformers.models.qwen3_5.modeling_qwen3_5"
+  },
+  "base_model_name_or_path": "",
+  "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,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.0,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "in_proj_qkv",
+    "k_proj",
+    "out_proj",
+    "q_proj",
+    "v_proj",
+    "o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": null,
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

package/checkpoint/adapter/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c857224bd3609c6a71eacf7f71dd021115fcc0f791936b1fc332e915b548a81
+size 61907944

package/checkpoint/head.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76e382f122abfa4e0c467d860a8d142d2fb6d2a98dc0ef9e19870bfc6eb296b4
+size 22221

package/checkpoint/model.json

@@ -0,0 +1,7 @@
+{
+  "model_id": "Qwen/Qwen3.8-27B",
+  "revision": "1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
+  "max_length": 4096,
+  "lora_rank": 8,
+  "method": "independent_candidate_lora_nll_brier"
+}

package/checkpoint/temperature.json

@@ -0,0 +1,6 @@
+{
+  "temperature": 2.5343690298472983,
+  "split": "calibration",
+  "n": 512,
+  "ids_sha256": "5c3dc72fe789a3e661a3b0d42b19e0b5483cf8e0df3727c999f1e01e764bf122"
+}

package/provenance.json

@@ -0,0 +1,388 @@
+{
+  "schema_version": 1,
+  "repo_id": "ZefanCai/Open-Jev-27B-v1.1",
+  "artifact_type": "qwen_lora_adapter_plus_scalar_decision_head",
+  "base_model": "Qwen/Qwen3.8-27B",
+  "base_revision": "1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
+  "base_weights_included": false,
+  "source_checkpoint_directory_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+  "source_checkpoint_files_sha256": {
+    "adapter/README.md": "f9e934433c6ae81516474d416c6b1dcf5193fe08bf41205363cf0b5b33760c1e",
+    "adapter/adapter_config.json": "0a11cc4a895a0ef163c1d25a4e2a5600c362adb805b4ce9e6278de9e8c78946a",
+    "adapter/adapter_model.safetensors": "1c857224bd3609c6a71eacf7f71dd021115fcc0f791936b1fc332e915b548a81",
+    "head.pt": "76e382f122abfa4e0c467d860a8d142d2fb6d2a98dc0ef9e19870bfc6eb296b4",
+    "model.json": "394c4ccc4f880172571abd141e16d00bf9dc140d1e252bae4e9185ee4cf8c508",
+    "temperature.json": "185c0b85539d195d02a2d4949295f0400fc208d9eb7ac4e5e6c97ebd339a235f"
+  },
+  "packaged_inference_directory_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+  "checkpoint_tree_preserved": true,
+  "packaging_omissions": [],
+  "training": {
+    "steps": 37160,
+    "max_length": 4096,
+    "lora_rank": 8,
+    "accumulation": 4,
+    "seed": 20260921,
+    "lr": 2e-05,
+    "head_lr": 5e-05,
+    "brier_weight": 0.1,
+    "training_sampling": "shuffled",
+    "commit": "7ac6bab261bd97efc3d446152999e94ae432e5b0",
+    "training_rows_consumed": 148640,
+    "resume_step": 28032,
+    "trainable_parameters": 15471617,
+    "identity_sha256": "9157ef46edc1177e59dd246c9e29142e87ee2545ccd4f591ae9222a1b2df4fe7",
+    "data_sha256": {
+      "train": "959dac64adc1717d92e9bb80371d4b940b2b5994f9e36e469b3eaed85883fdff",
+      "calibration": "6a02712bc6dce4c99cae3acfa54cf049553d12e4b795cfa4c170bd7de5732981",
+      "validation": "743ec57c076fee300bfe3e3eb364a8fb629addfd185724c0b66d58633bcd9372",
+      "test": "2cd250906060a570e0c596f7bb6ee7fb299fbf314e011a8c72ecb78d70248295",
+      "ood": "b575698e0f1006713ca736b96d1d8eb43a54d052e266c0f363b027b288f54498"
+    },
+    "data_manifest_sha256": "7b26f948d2ae11f20d7a18be437d1ada616fc479e1596ae94be7596787fe4e54",
+    "initialization_provenance": {
+      "kind": "ddp_snapshot_weights_only",
+      "source_manifest_sha256": "28aca3c8977191f976b0b961bd9924e4ba4ea511c6b627a887ae7cad937f41a5",
+      "source_training_state_sha256": "6ac6f647d4b6df7504e068b6f5408141f0f951c131b5f745272f94e0230b25d1",
+      "source_training_identity_sha256": "ad0ca316f7988163ba5ace364eea9d20104ad60e31bfaf2984a07bc013fb4a3f",
+      "source_distribution_sha256": "12559e06e61279e094226aded26834161eb6892ea6f42ae78f9defd8d07e22a1",
+      "source_completed_step": 26500,
+      "optimizer": "new_adamw_state",
+      "rng": "new_run_seed_and_rank_streams",
+      "step_cursor": 0,
+      "baseline_initialization": "warm_start_checkpoint"
+    },
+    "bitwise_ddp_continuation_claimed": false,
+    "dataset_name": "community-hard-mix-v2-final",
+    "unique_training_rows": 148639,
+    "wrapped_rows": 1,
+    "final_resume_manifest_sha256": "27cd0059093f13c36da66f0bdb2548c6b20aaeed45f72b046f08c1862fe77ba2",
+    "execution": {
+      "world_size": 4,
+      "backend": "nccl",
+      "device_type": "cuda",
+      "local_rows_per_step": 1,
+      "global_batch_size": 4,
+      "frozen_base": "BF16 FSDP2 per decoder layer and root; reshard_after_forward",
+      "trainables": "replicated FP32 LoRA/head; explicit gradient sum/4 before clipping",
+      "row_order": "train[(completed_steps * 4 + rank) % len(train)]",
+      "bitwise_ddp_equivalence": false,
+      "implementation_sha256": {
+        "train_shared.py": "d6cad3052e1fdf7476012ca290cadeb15cb3281b286b21e10bb9e73dd845cce7",
+        "sharded_model.py": "203ccce480104a3741058ad069ea243688d1e34aafdf6af214e64db7b6fb178b",
+        "shared_resources.py": "e910e6e556c8a4e1f6f404f7ae19ed8d5ad8bceb82e792f2b137498c4a989b48"
+      }
+    }
+  },
+  "calibration": {
+    "temperature": 2.5343690298472983,
+    "split": "calibration",
+    "n": 512,
+    "ids_sha256": "5c3dc72fe789a3e661a3b0d42b19e0b5483cf8e0df3727c999f1e01e764bf122"
+  },
+  "initialization": {
+    "kind": "weights_only_warm_start_from_previous_open_jev_27b",
+    "source_completed_step": 26500,
+    "source_dataset": "browser-drone-expansion-v1",
+    "source_data_manifest_sha256": "ba001b88787ce21576017897a0cbdedd5917ceb29ad3522b0e1fb7b4836d49df",
+    "source_snapshot_manifest_sha256": "28aca3c8977191f976b0b961bd9924e4ba4ea511c6b627a887ae7cad937f41a5",
+    "optimizer_for_new_mixture": "new AdamW state",
+    "new_mixture_step_cursor": 0
+  },
+  "public_data": {
+    "repo_id": "ZefanCai/Open-Jev-v1.1",
+    "repo_type": "dataset",
+    "config": "community-hard-mix-v2-redistributable",
+    "byte_identical_to_original_training_mixture": false,
+    "original_split_counts": {
+      "train": 148639,
+      "calibration": 26764,
+      "validation": 25482,
+      "test": 43301,
+      "ood": 84486
+    },
+    "public_split_counts": {
+      "train": 147139,
+      "calibration": 26675,
+      "validation": 25413,
+      "test": 43125,
+      "ood": 84267
+    },
+    "excluded_source": "wikispeedia-v1",
+    "excluded_training_rows": 1500,
+    "excluded_total_rows": 2053,
+    "reason": "Original Wikispeedia graph/path archive redistribution permission has not been confirmed."
+  },
+  "source_license_inventory": [
+    {
+      "component": "community-diversity-v2",
+      "manifest_sha256": "90a1f02ea457e65d3e3c40859fb73c5113e090fcc7dfe3cead72dbe6730bb2e1",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "wanli-decisions-v1",
+      "manifest_sha256": "38d50b03e62c63c561eff879351a73c426fec3a431b8cedd5edb0d564aa0f5af",
+      "declared_license": "CC-BY-4.0",
+      "note": null
+    },
+    {
+      "component": "browser-drone-expansion-v1",
+      "manifest_sha256": "ba001b88787ce21576017897a0cbdedd5917ceb29ad3522b0e1fb7b4836d49df",
+      "declared_license": null,
+      "note": "Aggregate source with per-component terms, including Wikispeedia with unconfirmed redistribution permission."
+    },
+    {
+      "component": "ir-control-v1",
+      "manifest_sha256": "5305dad7326fbf32aa006447137c5a3eb92872d90cde71302bc7efb39f882590",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "mailroom-control-v1",
+      "manifest_sha256": "afa9daabd5639a333c9adf8e89e7c7bbfcf84adfc051fb9843978c29703122f6",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "amount-extraction-control-v1",
+      "manifest_sha256": "6f89f81336ac36505c563193da27fca2155fa566175451854f0ac410979ea80b",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "entity-alignment-control-v1",
+      "manifest_sha256": "b3afddbeb926d57b7d7ea13c5fbda0fff878c647ea9ff1ee8a09b11d7f2f9eb5",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "email-selection-control-v1",
+      "manifest_sha256": "4c18cc791425003c5c8f052ce27ed034dfb54c79ddac24de13f7290ab3b3316c",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "phone-extraction-control-v1",
+      "manifest_sha256": "1001f1e04795308bf4f68db51076e8e1e1e47802986de54fc33b3b8d33548cfb",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "citation-control-v1",
+      "manifest_sha256": "069466d7536c579812bf2c870a9ea209d13a695005c299debffe61cd3aeffff4",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "sponsor-segment-control-v1",
+      "manifest_sha256": "6a143ede0485b22588fc92f4612f41428c46b1fa7d8f8cc0d32422d09b548f13",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "silent-failure-control-v1",
+      "manifest_sha256": "a7d326a480c0132dd608d0d7119a948bc16e36667e0ff7049e111f33c5445fac",
+      "declared_license": "CC0-1.0",
+      "note": null
+    },
+    {
+      "component": "context-retention-control-v1",
+      "manifest_sha256": "03badf69899809919e90bd299c5ddea4b3f8d646505d02f9544a1312e13bbd94",
+      "declared_license": "CC0-1.0",
+      "note": null
+    }
+  ],
+  "upstream_retrieval": {
+    "README.md": {
+      "url": "https://huggingface.co/Qwen/Qwen3.8-27B/raw/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0/README.md",
+      "sha256": "57e4bdb258ee1a7d2635c5174ebd4e56abe392505cdb5f8bbb356b0dc4293641",
+      "bytes": 65012
+    },
+    "LICENSE": {
+      "url": "https://huggingface.co/Qwen/Qwen3.8-27B/raw/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0/LICENSE",
+      "sha256": "bbedc3fda3305820b977265f01b8619d87570a6739de3a5582c3464840f1e57a",
+      "bytes": 11544
+    },
+    "config.json": {
+      "url": "https://huggingface.co/Qwen/Qwen3.8-27B/raw/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0/config.json",
+      "sha256": "191e0af232104ed8b65258cf3fb2b842e288008baca7633c11b82a1ac7203aab",
+      "bytes": 4312
+    },
+    "revision-api.json": {
+      "url": "https://huggingface.co/api/models/Qwen/Qwen3.8-27B/revision/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
+      "sha256": "18803d62aeff63743add15cff5e7b6c8746a1df44b8006ba3070599824e0944f",
+      "bytes": 24096
+    }
+  },
+  "evaluation": {
+    "full_internal": {
+      "status": "passed_complete_full24_audit",
+      "publishable": true,
+      "report": "verification/internal-full.json",
+      "report_sha256": "19a154c1a6ed3935bac266ab04b8a395bc180739540127b4b3802190e3c34f97",
+      "independent_audit_sha256": "650642bac1c7440f7cdbaf044068b8870ef7a514510dfd8a6cb01756a9304795",
+      "campaign_manifest_sha256": "491a4ee96d1ef6509ead733daffdf2318f3bb43d9d5bf38715f2cd0be8a0a9d4",
+      "evaluation_code_commit": "9eebd42de6137735fc5fd85b499aa1ffa256426d",
+      "checkpoint_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+      "coverage": {
+        "total": 127787,
+        "successful": 127787,
+        "failed": 0,
+        "missing": 0,
+        "duplicates": 0,
+        "imported": 4491,
+        "new": 123296,
+        "completed_groups": 6,
+        "owned_processes_remaining": 0
+      },
+      "panels": {
+        "old_test": {
+          "count": 10532,
+          "hard_count": 10046,
+          "soft_count": 486,
+          "hard_correct": 9876,
+          "accuracy": 0.9830778419271352,
+          "expected_accuracy": 0.954681027079432,
+          "nll": 0.13386432263131773,
+          "brier": 0.03458594885426232,
+          "expected_brier": 0.06263691316614574,
+          "multiclass_ece": 0.029161848979411622,
+          "ece_bins": 15
+        },
+        "expanded_test": {
+          "count": 43301,
+          "hard_count": 42789,
+          "soft_count": 512,
+          "hard_correct": 41357,
+          "accuracy": 0.966533454859894,
+          "expected_accuracy": 0.9595090315974361,
+          "nll": 0.12983825446990038,
+          "brier": 0.05746167254927552,
+          "expected_brier": 0.06460776054107147,
+          "multiclass_ece": 0.037463708716989785,
+          "ece_bins": 15
+        },
+        "old_ood": {
+          "count": 15920,
+          "hard_count": 15446,
+          "soft_count": 474,
+          "hard_correct": 14825,
+          "accuracy": 0.9597954162890069,
+          "expected_accuracy": 0.9430203212128463,
+          "nll": 0.15887220783486702,
+          "brier": 0.06296576556098005,
+          "expected_brier": 0.08029650520798853,
+          "multiclass_ece": 0.024102283529718986,
+          "ece_bins": 15
+        },
+        "expanded_ood": {
+          "count": 84486,
+          "hard_count": 83924,
+          "soft_count": 562,
+          "hard_correct": 80934,
+          "accuracy": 0.9643725275249034,
+          "expected_accuracy": 0.9606666609107841,
+          "nll": 0.13434172410803066,
+          "brier": 0.060333553335607484,
+          "expected_brier": 0.06415555195289763,
+          "multiclass_ece": 0.01820277638023621,
+          "ece_bins": 15
+        }
+      },
+      "protocol": {
+        "max_length": 4096,
+        "truncation": false,
+        "cache": false,
+        "row_batch_size": 1,
+        "temperature_fitted_on_evaluation": false,
+        "base_dtype": "BF16",
+        "head_dtype": "FP32",
+        "hard_accuracy": "Exact one-hot targets only; all rows contribute to probability metrics.",
+        "subset": "Old Test/OOD are content-identical subsets of expanded Test/OOD; panels overlap."
+      }
+    },
+    "jevbench": {
+      "status": "complete_independently_audited",
+      "report": "verification/jevbench.json",
+      "report_sha256": "9a78ffb3aecb113e26bf4d83fd5ad7467141e3bdf84535e90ba71d634ec2a85d",
+      "scope": "231 public tasks only, including 111 Hard; not the complete 534-task benchmark.",
+      "checkpoint_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+      "max_length": 16384,
+      "temperature": 2.5343690298472983,
+      "overall": {
+        "n_planned": 231,
+        "n_attempted": 231,
+        "n_scorable": 231,
+        "n_valid": 231,
+        "n_correct": 197,
+        "accuracy": 0.8528138528138528,
+        "schema_validity": 1.0,
+        "schema_validity_strict": 1.0,
+        "n_renormalized": 0,
+        "operational_success": 1.0,
+        "brier_mean": 0.2419831167860275,
+        "ordinal_mae": 0.3798003972534468
+      },
+      "hard": {
+        "n_planned": 111,
+        "n_attempted": 111,
+        "n_valid": 111,
+        "n_correct": 80,
+        "accuracy": 0.7207207207207207
+      },
+      "comparison_baselines": [
+        {
+          "label": "Released Open-Jev 2B",
+          "overall": {
+            "n_correct": 150,
+            "n_planned": 231,
+            "accuracy": 0.6493506493506493
+          },
+          "hard": {
+            "n_correct": 46,
+            "n_planned": 111,
+            "accuracy": 0.4144144144144144
+          }
+        },
+        {
+          "label": "Released Open-Jev 9B",
+          "overall": {
+            "n_correct": 179,
+            "n_planned": 231,
+            "accuracy": 0.7748917748917749
+          },
+          "hard": {
+            "n_correct": 66,
+            "n_planned": 111,
+            "accuracy": 0.5945945945945946
+          }
+        },
+        {
+          "label": "Jev 1.13.0",
+          "overall": {
+            "n_correct": 200,
+            "n_planned": 231,
+            "accuracy": 0.8658008658008658
+          },
+          "hard": {
+            "n_correct": 81,
+            "n_planned": 111,
+            "accuracy": 0.7297297297297297
+          }
+        }
+      ]
+    },
+    "expected_original_test_rows": 43301,
+    "expected_original_ood_rows": 84486,
+    "latency_claims": null,
+    "sampled_training_metrics_included": false
+  },
+  "training_run_artifact_hashes": {
+    "run.json": "15721f8b41a7dd4e6a829c7fe06cec1e609f109169598edd1441bd79308a4e1b",
+    "summary.json": "c2870c2518100402c96b1ce1a7edec7f5a1afe5b322f37776f1477b5a8a3a8e6",
+    "reload_check.jsonl": "f84fa6b2bf418670301b12dac60cfddea4c41fe19be32e9bae57b9cadb2ecf07",
+    "calibration.jsonl": "2d9b457512d1b470d4c2c756fcba68d6c08bd0725df935e7096434b5a23164ae"
+  },
+  "scope": "Source hashes bind private training records without distributing raw records, prompts, targets or credentials."
+}

provenance/lexical-overlap-screen.json

@@ -0,0 +1,33 @@
+{
+  "status": "no_detected_overlap",
+  "request_files": [
+    {
+      "sha256": "2e20ff5f94dd04d015b3a7b9abd955757b6c5be25b55b97fcd141cb52f88b4aa",
+      "name": "requests.json"
+    },
+    {
+      "sha256": "2a373b30cdbcd26fdbc538252c7a5d1c5250f86160a57e9a96d70ca2f2d6106c",
+      "name": "requests.json"
+    }
+  ],
+  "request_count": 531,
+  "dataset_files": [
+    {
+      "path": "data/community-hard-mix-v2-final/train.jsonl",
+      "sha256": "959dac64adc1717d92e9bb80371d4b940b2b5994f9e36e469b3eaed85883fdff"
+    }
+  ],
+  "training_rows": {
+    "data/community-hard-mix-v2-final": 148639
+  },
+  "matched_rows": 0,
+  "excluded_groups": [],
+  "matches": [],
+  "policy": "Visible state/question/options only; no targets, gold, answers or rationales read from benchmark requests.",
+  "limits": [
+    "Lexical screen cannot detect paraphrases or prove semantic independence.",
+    "Does not assess foundation-model pretraining exposure.",
+    "13-token overlap may also flag harmless shared wording; matching groups are quarantined conservatively.",
+    "Shorter state leaves are covered by full-input exact matching only."
+  ]
+}

provenance/original-training-mixture-manifest.json

@@ -0,0 +1,488 @@
+{
+  "schema_version": 1,
+  "configuration": {
+    "version": "community-hard-mix-v2",
+    "seed": 20260921,
+    "replay_cap_per_source": 1500,
+    "input_bindings": [
+      {
+        "category": "new",
+        "path": "data/community-diversity-v2",
+        "manifest_sha256": "90a1f02ea457e65d3e3c40859fb73c5113e090fcc7dfe3cead72dbe6730bb2e1",
+        "files_sha256": {
+          "train": "4274569425c7003a8705f804b27402d6e0963167c7a21dd8c491573b40520f16",
+          "calibration": "ed07c44e9d2793f5e4f3449e48152ffe3915a7fed89edac16f4cd7861ccb3cd7",
+          "validation": "b12851e59eb3eee5d917baddde35ab7bfc89910f77d91f4a554937ec59edf833",
+          "test": "71168841444f9d478e6bbca6148a46a4dc4426f12f283205d6a02517145f7a6d",
+          "ood": "4359ec2a43d0d17d7b4f8059954af648431cff16eb4cfc97c27894434c26c49a"
+        }
+      },
+      {
+        "category": "new",
+        "path": "data/wanli-decisions-v1",
+        "manifest_sha256": "38d50b03e62c63c561eff879351a73c426fec3a431b8cedd5edb0d564aa0f5af",
+        "files_sha256": {
+          "train": "62b2561f1a43d157eb854916a4d24047f78e61dddb2e700b49d3268111449a85",
+          "calibration": "7253def87adf070728e534f39284df1e349e30a13c1d65291d5189afc7f0aa70",
+          "validation": "601ced227bb054b035654073c74eb4bafef29b3e368c325b083dab5c24fae384",
+          "test": "5383a944978701468b8bdc884ed7ad41e78f2aea369317feb08b931af865de4f",
+          "ood": "ce1fbebd84b6ca95d3dd94b2e93b224412a78090d9cafc2d4d560e61690aa13e"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/browser-drone-expansion-v1",
+        "manifest_sha256": "ba001b88787ce21576017897a0cbdedd5917ceb29ad3522b0e1fb7b4836d49df",
+        "files_sha256": {
+          "train": "91a3e3c715abd735509a5a23453161475aff85626f2c47563a8fab4e7dc79239",
+          "calibration": "8a366dbb71dbe57b98a25da3071fc858beb7e8567adf4abf952c3d98f79bb94c",
+          "validation": "bb49a3ca608dbeb543c56a20ee2f32e75f65be835fcb40ec40014af7351c6c7d",
+          "test": "662d5e79da70eba77ad740eb37576b8157e53e5c1d8cd9344c60ec56d3969637",
+          "ood": "364992ae82a2d8b6f3f4ed68a0efa06d9a6b9ae469d81113d4165f762eec0a94"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/ir-control-v1",
+        "manifest_sha256": "5305dad7326fbf32aa006447137c5a3eb92872d90cde71302bc7efb39f882590",
+        "files_sha256": {
+          "train": "6d33876576fb7989abbf9edcd878074f4da7d4f502d770d110a848bbe259541d",
+          "calibration": "7ad15592ec8b1d2e6c2d819fb186dc2dec47f8ed9214072f6f45452be7194e61",
+          "validation": "0ecbe867ae894a83101805328be1b1d983cc210fb624fa78d2aedd59f8b702ec",
+          "test": "97637e12062c98447b49cfa8cc9669e86525dce0078657a08fd9f1b54e5e2869",
+          "ood": "fa6332b13395ef37ab4c88e4c5888201f8c814fdd26d387bc5684e7bf410076b"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/mailroom-control-v1",
+        "manifest_sha256": "afa9daabd5639a333c9adf8e89e7c7bbfcf84adfc051fb9843978c29703122f6",
+        "files_sha256": {
+          "train": "5a19ebf41125093c8bb7aa0861842673686014b2c288e79bfbb54a0ea1a4a86d",
+          "calibration": "9765f7481ff6eda49733d1cdee648c771bb1a2055bcc7df9bac5d2acda9b592b",
+          "validation": "97558afbd86cb84cada925c9513fa868db3e47db83e51fc1dafe955240d05516",
+          "test": "9bb9976cfa9e3e5dc084b86d9cbeb0bc201465a5ca0249af6781c02bdf635938",
+          "ood": "5667325039486c28ade0b752c3fbf694aa7c73e7998a274744b09b5dea207d0e"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/amount-extraction-control-v1",
+        "manifest_sha256": "6f89f81336ac36505c563193da27fca2155fa566175451854f0ac410979ea80b",
+        "files_sha256": {
+          "train": "68a34ca1371b60dc9b49c7d26882893758c16bc1dbb85a636cf0643ed7f5047d",
+          "calibration": "4db65ea37fc9569f3666d87762aec37c6786dfa7e7caee9dc70ef6e1420d6e6b",
+          "validation": "b9bc54dd4c39eb1ab40313bf0ce1339c48b579077000b5de1c0a9ffaba802e76",
+          "test": "f8500ef9dbd65280b2c58edcfc579dd175ddde52bc2b19b6b97c5cd84bff8bd8",
+          "ood": "11f3348654087a561fc3ec934144712e38decf6cfb43c21efb5b96099e0f3e93"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/entity-alignment-control-v1",
+        "manifest_sha256": "b3afddbeb926d57b7d7ea13c5fbda0fff878c647ea9ff1ee8a09b11d7f2f9eb5",
+        "files_sha256": {
+          "train": "ff4f67f3425ce6568388c44524f04fce438c580d5eb3f4bb0a8f5508f11e4234",
+          "calibration": "dd6abd73a852a6067bca90bb9f378b0eae8acb5cfcaf8df71473c8f64494c01f",
+          "validation": "e0a3d143d6dc4a0e983124746bb02537ab200a322aba84a176daea499c75e2b6",
+          "test": "ae101aa7d0692106f71634ccebf39b7062cb37181025af9b9b8955b0a82a0048",
+          "ood": "8c4172dc683c41caf7355291b1bd9e27aa558b7073037bde9a87dd3a5cc56359"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/email-selection-control-v1",
+        "manifest_sha256": "4c18cc791425003c5c8f052ce27ed034dfb54c79ddac24de13f7290ab3b3316c",
+        "files_sha256": {
+          "train": "2127d52a792e1e7d2727315aaf7368e61b50d9853559f4e694aa96964e3fe492",
+          "calibration": "d6a6ddb37d4d086028a8c68f6b2c677ff4860ac0441c35e586d4f2233aa4bdd9",
+          "validation": "575c2a53df3b075210cb6f4845e94a15123321af0945208b0b7eba00cb044ad9",
+          "test": "9d4138f70de5473700dd59a92bb893d87e68d98ce0a76ead98c907fe8256e28e",
+          "ood": "975ec0cff2afd7188313682b73de64be76ad7f8fdc42603fefb1c2cc616ecd04"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/phone-extraction-control-v1",
+        "manifest_sha256": "1001f1e04795308bf4f68db51076e8e1e1e47802986de54fc33b3b8d33548cfb",
+        "files_sha256": {
+          "train": "407bd4c4d0b9e703d83e2db6327acb2604f032188a0535578bce7d9d4688859e",
+          "calibration": "48a9d3aa1da99f6a8794e5a139e9b3d3a2b6f127edfab478e2f00d2d6c119dca",
+          "validation": "7a078bb9b2bd0b4c58858ace347ad4a9c513122b628799193c1214b5ff6ff395",
+          "test": "dfc1180917d4c8717bc15c5c2162e2dbea24172158d593b062c8cb5eb1360903",
+          "ood": "576fb66309089809a4646aeb88e3cc3d187c42498eb6d564c9bc79cad730f88b"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/citation-control-v1",
+        "manifest_sha256": "069466d7536c579812bf2c870a9ea209d13a695005c299debffe61cd3aeffff4",
+        "files_sha256": {
+          "train": "c8e939ea7e49290a8d551b93cf759dda68dd25bce264dbd0461cf2fb1e7ca448",
+          "calibration": "feae8a0adb34690d0dadb920213b3da3f29b52aaf4c938bd3279c58821a69e04",
+          "validation": "e3b7dd9759199d9aa64e1df94b1efdd0ef91a7fdc7090ce0aa6943604fefb1f7",
+          "test": "fd55095c6387a3027e0eb9639e3c6a802cdff29e587ef3118457d84285016a50",
+          "ood": "f85863a39afaf12cb5ab3e4ab554f9bba93a41b1741629daa6a855c9dd3520eb"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/sponsor-segment-control-v1",
+        "manifest_sha256": "6a143ede0485b22588fc92f4612f41428c46b1fa7d8f8cc0d32422d09b548f13",
+        "files_sha256": {
+          "train": "cd8a28b09840594f945d389447e77f0d216c852028d4cce05cda28f2506f7148",
+          "calibration": "79540596a197fc4364fb3fe4c7525ad8347ee72e48fec3cbfa0b43e5b47b3322",
+          "validation": "12a503bb310414659469ddf14aede84b5a64f4dc166bcbf28d870b9fb3a01c60",
+          "test": "642f6e6577a4761dee2b862f914978bd6fc6697979e1a954fd9eb9e163865a17",
+          "ood": "29ad7566e7b4989c7cec7aee6d8266db26d19a0829d43a5e179cdd2fade325d9"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/silent-failure-control-v1",
+        "manifest_sha256": "a7d326a480c0132dd608d0d7119a948bc16e36667e0ff7049e111f33c5445fac",
+        "files_sha256": {
+          "train": "3b83b1d937ca2df42e94d15a7ef70125672f653ceeb82fe3790a7b07d2bf8a8f",
+          "calibration": "2c86b1ddc82f20fe6979920620caa5dbb3568dcc159c307ee32b49e99bc2c85d",
+          "validation": "173e5a00b1a5573d91ea5ba323bea2b1d4358eea44cceee85dce5823d9ea221a",
+          "test": "37c95060bd0c38d9dd3315e30d60ad69a38d8fe717fca32e7c77b4ed4a944024",
+          "ood": "3fd69c6e86ab7e6b6034e4b48bd55d845488efca1815626af9b9103e30cfd299"
+        }
+      },
+      {
+        "category": "replay",
+        "path": "data/context-retention-control-v1",
+        "manifest_sha256": "03badf69899809919e90bd299c5ddea4b3f8d646505d02f9544a1312e13bbd94",
+        "files_sha256": {
+          "train": "d6c4cdc5572da328675b9e3c5eb8af8ef186a9a3fbd5c3bdeaec4a7be21bca06",
+          "calibration": "b157cd03922bd89056ca3678223c1f64bd152d81d5ba4c14b1ec0088f4f4d2ca",
+          "validation": "20063011a0fbdc07582ce8da7e7f40502995d03460e73d7ac902a23bb32bc7cf",
+          "test": "a5e3adc7dbb90209ef4ffd657bcb004424ab71807bdf076ae234626cb769b3ee",
+          "ood": "558db34ab2be4c12fffa611a47612506977245e2d4c598caa561c374c7f68b97"
+        }
+      }
+    ],
+    "replay_policy": "Hash-ordered round robin across parent groups; no duplication or oversampling.",
+    "new_source_policy": "One full pass over retained rows; four-rank training wraps at most three rows to complete its final step.",
+    "excluded_groups": [],
+    "filtered": {
+      "replay_train_cap": 232619
+    }
+  },
+  "summary": {
+    "records": 328672,
+    "groups": 76619,
+    "unique_inputs": 327954,
+    "splits": {
+      "calibration": 26764,
+      "ood": 84486,
+      "test": 43301,
+      "train": 148639,
+      "validation": 25482
+    },
+    "kinds": {
+      "choice": 185660,
+      "noul": 114562,
+      "score": 28450
+    },
+    "families": {
+      "doom_basic": 4239,
+      "evidence": 112873,
+      "natural_language_inference": 101207,
+      "policy": 72344,
+      "routing": 17720,
+      "rubric": 18236,
+      "wikiracing": 2053
+    },
+    "sources": {
+      "amount-extraction-control-v1": 18116,
+      "browser-control-v1": 9666,
+      "citation-control-v1": 2980,
+      "community-diversity-v2/browser_tools/argument_binding": 640,
+      "community-diversity-v2/browser_tools/confirmation_gate": 640,
+      "community-diversity-v2/browser_tools/date_range_form": 640,
+      "community-diversity-v2/browser_tools/goal_completion": 640,
+      "community-diversity-v2/browser_tools/idempotent_retry": 640,
+      "community-diversity-v2/browser_tools/pagination": 640,
+      "community-diversity-v2/browser_tools/stale_dom": 640,
+      "community-diversity-v2/browser_tools/tool_scope": 640,
+      "community-diversity-v2/browser_tools/typing_fallback": 640,
+      "community-diversity-v2/browser_tools/upload_constraint": 640,
+      "community-diversity-v2/contracts/assignment_consent": 640,
+      "community-diversity-v2/contracts/audit_right": 640,
+      "community-diversity-v2/contracts/automatic_renewal": 640,
+      "community-diversity-v2/contracts/change_of_control": 640,
+      "community-diversity-v2/contracts/confidentiality_survival": 640,
+      "community-diversity-v2/contracts/exclusivity_exception": 640,
+      "community-diversity-v2/contracts/governing_law_venue": 640,
+      "community-diversity-v2/contracts/liability_cap": 640,
+      "community-diversity-v2/contracts/noncompete_scope": 640,
+      "community-diversity-v2/contracts/termination_cure": 640,
+      "community-diversity-v2/games/crafting_dependencies": 640,
+      "community-diversity-v2/games/escort_sync": 640,
+      "community-diversity-v2/games/health_pickup": 640,
+      "community-diversity-v2/games/pit_stop": 640,
+      "community-diversity-v2/games/projectile_dodge": 640,
+      "community-diversity-v2/games/racing_braking": 640,
+      "community-diversity-v2/games/racing_overtake": 640,
+      "community-diversity-v2/games/reload_cover": 640,
+      "community-diversity-v2/games/safe_path": 640,
+      "community-diversity-v2/games/weapon_selection": 640,
+      "community-diversity-v2/ir/access_schedule": 640,
+      "community-diversity-v2/ir/api_auth_migration": 640,
+      "community-diversity-v2/ir/connecting_route": 640,
+      "community-diversity-v2/ir/hardware_manual": 640,
+      "community-diversity-v2/ir/incident_runbook": 640,
+      "community-diversity-v2/ir/inventory_answer": 640,
+      "community-diversity-v2/ir/regional_returns": 640,
+      "community-diversity-v2/ir/release_migration": 640,
+      "community-diversity-v2/ir/study_population": 640,
+      "community-diversity-v2/ir/warranty_transfer": 640,
+      "community-diversity-v2/rag/citation_coverage": 640,
+      "community-diversity-v2/rag/contradiction_resolution": 640,
+      "community-diversity-v2/rag/entity_alias_chain": 640,
+      "community-diversity-v2/rag/independent_sources": 640,
+      "community-diversity-v2/rag/negative_evidence": 640,
+      "community-diversity-v2/rag/numerical_claim": 640,
+      "community-diversity-v2/rag/policy_exception": 640,
+      "community-diversity-v2/rag/qualified_comparison": 640,
+      "community-diversity-v2/rag/temporal_fact": 640,
+      "community-diversity-v2/rag/two_hop_ownership": 640,
+      "community-diversity-v2/rubric_judge/abstention_quality": 640,
+      "community-diversity-v2/rubric_judge/balanced_tradeoff": 640,
+      "community-diversity-v2/rubric_judge/citation_attribution": 640,
+      "community-diversity-v2/rubric_judge/comparison_fairness": 640,
+      "community-diversity-v2/rubric_judge/faithful_summary": 640,
+      "community-diversity-v2/rubric_judge/format_and_content": 640,
+      "community-diversity-v2/rubric_judge/instruction_priority": 640,
+      "community-diversity-v2/rubric_judge/negation_preservation": 640,
+      "community-diversity-v2/rubric_judge/trace_consistency": 640,
+      "community-diversity-v2/rubric_judge/unit_conversion": 640,
+      "community-diversity-v2/shell_history/archive_extract": 640,
+      "community-diversity-v2/shell_history/checksum_verify": 640,
+      "community-diversity-v2/shell_history/dependency_install": 640,
+      "community-diversity-v2/shell_history/exact_repo_test": 640,
+      "community-diversity-v2/shell_history/git_history_query": 640,
+      "community-diversity-v2/shell_history/incremental_copy": 640,
+      "community-diversity-v2/shell_history/log_time_window": 640,
+      "community-diversity-v2/shell_history/port_inspection": 640,
+      "community-diversity-v2/shell_history/quoted_filename": 640,
+      "community-diversity-v2/shell_history/search_scope": 640,
+      "community-diversity-v2/support/account_recovery": 640,
+      "community-diversity-v2/support/damaged_order": 640,
+      "community-diversity-v2/support/delayed_parcel": 640,
+      "community-diversity-v2/support/duplicate_charge": 640,
+      "community-diversity-v2/support/outage_compensation": 640,
+      "community-diversity-v2/support/partial_return": 640,
+      "community-diversity-v2/support/privacy_export": 640,
+      "community-diversity-v2/support/refund_window": 640,
+      "community-diversity-v2/support/subscription_cancellation": 640,
+      "community-diversity-v2/support/warranty_repair": 640,
+      "context-retention-control-v1": 5196,
+      "customer-control-v1": 3294,
+      "drone-control-v1": 11055,
+      "email-selection-control-v1": 3282,
+      "entity-alignment-control-v1": 5756,
+      "ir-control-v1": 5734,
+      "mailroom-control-v1": 42828,
+      "painting-geometry-v1": 12508,
+      "phone-extraction-control-v1": 8150,
+      "reasoning-control-v1": 3558,
+      "silent-failure-control-v1": 4668,
+      "snake-v1": 7876,
+      "sponsor-segment-control-v1": 5955,
+      "tic_tac_toe-v1": 2534,
+      "tile_platformer-v1": 1697,
+      "trex_runner-v1": 36,
+      "vizdoom-basic-v1": 4239,
+      "wanli-decisions-v1": 101207,
+      "wikispeedia-v1": 2053,
+      "workflow-controls-v1/agent_trace_observability": 2970,
+      "workflow-controls-v1/customer_service": 3108,
+      "workflow-controls-v1/invoice_processing": 3630,
+      "workflow-controls-v1/security_incidents": 5376
+    }
+  },
+  "files_sha256": {
+    "train.jsonl": "959dac64adc1717d92e9bb80371d4b940b2b5994f9e36e469b3eaed85883fdff",
+    "calibration.jsonl": "6a02712bc6dce4c99cae3acfa54cf049553d12e4b795cfa4c170bd7de5732981",
+    "validation.jsonl": "743ec57c076fee300bfe3e3eb364a8fb629addfd185724c0b66d58633bcd9372",
+    "test.jsonl": "2cd250906060a570e0c596f7bb6ee7fb299fbf314e011a8c72ecb78d70248295",
+    "ood.jsonl": "b575698e0f1006713ca736b96d1d8eb43a54d052e266c0f363b027b288f54498"
+  },
+  "model_input_fields": [
+    "state",
+    "question",
+    "kind",
+    "options"
+  ],
+  "counts": {
+    "train": 148639,
+    "calibration": 26764,
+    "validation": 25482,
+    "test": 43301,
+    "ood": 84486
+  },
+  "sha256": {
+    "train": "959dac64adc1717d92e9bb80371d4b940b2b5994f9e36e469b3eaed85883fdff",
+    "calibration": "6a02712bc6dce4c99cae3acfa54cf049553d12e4b795cfa4c170bd7de5732981",
+    "validation": "743ec57c076fee300bfe3e3eb364a8fb629addfd185724c0b66d58633bcd9372",
+    "test": "2cd250906060a570e0c596f7bb6ee7fb299fbf314e011a8c72ecb78d70248295",
+    "ood": "b575698e0f1006713ca736b96d1d8eb43a54d052e266c0f363b027b288f54498"
+  },
+  "training_by_source": {
+    "amount-extraction-control-v1": 1500,
+    "browser-control-v1": 1500,
+    "citation-control-v1": 1500,
+    "community-diversity-v2/browser_tools/date_range_form": 640,
+    "community-diversity-v2/browser_tools/pagination": 640,
+    "community-diversity-v2/browser_tools/stale_dom": 640,
+    "community-diversity-v2/browser_tools/tool_scope": 640,
+    "community-diversity-v2/browser_tools/typing_fallback": 640,
+    "community-diversity-v2/browser_tools/upload_constraint": 640,
+    "community-diversity-v2/contracts/assignment_consent": 640,
+    "community-diversity-v2/contracts/audit_right": 640,
+    "community-diversity-v2/contracts/automatic_renewal": 640,
+    "community-diversity-v2/contracts/confidentiality_survival": 640,
+    "community-diversity-v2/contracts/liability_cap": 640,
+    "community-diversity-v2/contracts/noncompete_scope": 640,
+    "community-diversity-v2/games/escort_sync": 640,
+    "community-diversity-v2/games/health_pickup": 640,
+    "community-diversity-v2/games/projectile_dodge": 640,
+    "community-diversity-v2/games/reload_cover": 640,
+    "community-diversity-v2/games/safe_path": 640,
+    "community-diversity-v2/games/weapon_selection": 640,
+    "community-diversity-v2/ir/connecting_route": 640,
+    "community-diversity-v2/ir/hardware_manual": 640,
+    "community-diversity-v2/ir/inventory_answer": 640,
+    "community-diversity-v2/ir/regional_returns": 640,
+    "community-diversity-v2/ir/release_migration": 640,
+    "community-diversity-v2/ir/study_population": 640,
+    "community-diversity-v2/rag/citation_coverage": 640,
+    "community-diversity-v2/rag/contradiction_resolution": 640,
+    "community-diversity-v2/rag/negative_evidence": 640,
+    "community-diversity-v2/rag/numerical_claim": 640,
+    "community-diversity-v2/rag/temporal_fact": 640,
+    "community-diversity-v2/rag/two_hop_ownership": 640,
+    "community-diversity-v2/rubric_judge/abstention_quality": 640,
+    "community-diversity-v2/rubric_judge/comparison_fairness": 640,
+    "community-diversity-v2/rubric_judge/faithful_summary": 640,
+    "community-diversity-v2/rubric_judge/format_and_content": 640,
+    "community-diversity-v2/rubric_judge/negation_preservation": 640,
+    "community-diversity-v2/rubric_judge/unit_conversion": 640,
+    "community-diversity-v2/shell_history/archive_extract": 640,
+    "community-diversity-v2/shell_history/dependency_install": 640,
+    "community-diversity-v2/shell_history/exact_repo_test": 640,
+    "community-diversity-v2/shell_history/git_history_query": 640,
+    "community-diversity-v2/shell_history/quoted_filename": 640,
+    "community-diversity-v2/shell_history/search_scope": 640,
+    "community-diversity-v2/support/account_recovery": 640,
+    "community-diversity-v2/support/delayed_parcel": 640,
+    "community-diversity-v2/support/duplicate_charge": 640,
+    "community-diversity-v2/support/refund_window": 640,
+    "community-diversity-v2/support/subscription_cancellation": 640,
+    "community-diversity-v2/support/warranty_repair": 640,
+    "context-retention-control-v1": 1500,
+    "customer-control-v1": 1500,
+    "drone-control-v1": 1500,
+    "email-selection-control-v1": 1500,
+    "entity-alignment-control-v1": 1500,
+    "ir-control-v1": 1500,
+    "mailroom-control-v1": 1500,
+    "painting-geometry-v1": 1500,
+    "phone-extraction-control-v1": 1500,
+    "reasoning-control-v1": 1500,
+    "silent-failure-control-v1": 1500,
+    "snake-v1": 1500,
+    "sponsor-segment-control-v1": 1500,
+    "tic_tac_toe-v1": 1500,
+    "tile_platformer-v1": 1346,
+    "trex_runner-v1": 28,
+    "vizdoom-basic-v1": 1500,
+    "wanli-decisions-v1": 82045,
+    "wikispeedia-v1": 1500,
+    "workflow-controls-v1/agent_trace_observability": 1500,
+    "workflow-controls-v1/customer_service": 1500,
+    "workflow-controls-v1/invoice_processing": 1500,
+    "workflow-controls-v1/security_incidents": 1500
+  },
+  "training_groups_by_source": {
+    "amount-extraction-control-v1": 133,
+    "browser-control-v1": 631,
+    "citation-control-v1": 126,
+    "community-diversity-v2/browser_tools/date_range_form": 80,
+    "community-diversity-v2/browser_tools/pagination": 80,
+    "community-diversity-v2/browser_tools/stale_dom": 80,
+    "community-diversity-v2/browser_tools/tool_scope": 80,
+    "community-diversity-v2/browser_tools/typing_fallback": 80,
+    "community-diversity-v2/browser_tools/upload_constraint": 80,
+    "community-diversity-v2/contracts/assignment_consent": 80,
+    "community-diversity-v2/contracts/audit_right": 80,
+    "community-diversity-v2/contracts/automatic_renewal": 80,
+    "community-diversity-v2/contracts/confidentiality_survival": 80,
+    "community-diversity-v2/contracts/liability_cap": 80,
+    "community-diversity-v2/contracts/noncompete_scope": 80,
+    "community-diversity-v2/games/escort_sync": 80,
+    "community-diversity-v2/games/health_pickup": 80,
+    "community-diversity-v2/games/projectile_dodge": 80,
+    "community-diversity-v2/games/reload_cover": 80,
+    "community-diversity-v2/games/safe_path": 80,
+    "community-diversity-v2/games/weapon_selection": 80,
+    "community-diversity-v2/ir/connecting_route": 80,
+    "community-diversity-v2/ir/hardware_manual": 80,
+    "community-diversity-v2/ir/inventory_answer": 80,
+    "community-diversity-v2/ir/regional_returns": 80,
+    "community-diversity-v2/ir/release_migration": 80,
+    "community-diversity-v2/ir/study_population": 80,
+    "community-diversity-v2/rag/citation_coverage": 80,
+    "community-diversity-v2/rag/contradiction_resolution": 80,
+    "community-diversity-v2/rag/negative_evidence": 80,
+    "community-diversity-v2/rag/numerical_claim": 80,
+    "community-diversity-v2/rag/temporal_fact": 80,
+    "community-diversity-v2/rag/two_hop_ownership": 80,
+    "community-diversity-v2/rubric_judge/abstention_quality": 80,
+    "community-diversity-v2/rubric_judge/comparison_fairness": 80,
+    "community-diversity-v2/rubric_judge/faithful_summary": 80,
+    "community-diversity-v2/rubric_judge/format_and_content": 80,
+    "community-diversity-v2/rubric_judge/negation_preservation": 80,
+    "community-diversity-v2/rubric_judge/unit_conversion": 80,
+    "community-diversity-v2/shell_history/archive_extract": 80,
+    "community-diversity-v2/shell_history/dependency_install": 80,
+    "community-diversity-v2/shell_history/exact_repo_test": 80,
+    "community-diversity-v2/shell_history/git_history_query": 80,
+    "community-diversity-v2/shell_history/quoted_filename": 80,
+    "community-diversity-v2/shell_history/search_scope": 80,
+    "community-diversity-v2/support/account_recovery": 80,
+    "community-diversity-v2/support/delayed_parcel": 80,
+    "community-diversity-v2/support/duplicate_charge": 80,
+    "community-diversity-v2/support/refund_window": 80,
+    "community-diversity-v2/support/subscription_cancellation": 80,
+    "community-diversity-v2/support/warranty_repair": 80,
+    "context-retention-control-v1": 255,
+    "customer-control-v1": 701,
+    "drone-control-v1": 311,
+    "email-selection-control-v1": 134,
+    "entity-alignment-control-v1": 124,
+    "ir-control-v1": 127,
+    "mailroom-control-v1": 256,
+    "painting-geometry-v1": 46,
+    "phone-extraction-control-v1": 130,
+    "reasoning-control-v1": 1500,
+    "silent-failure-control-v1": 268,
+    "snake-v1": 1500,
+    "sponsor-segment-control-v1": 235,
+    "tic_tac_toe-v1": 452,
+    "tile_platformer-v1": 63,
+    "trex_runner-v1": 28,
+    "vizdoom-basic-v1": 309,
+    "wanli-decisions-v1": 44541,
+    "wikispeedia-v1": 224,
+    "workflow-controls-v1/agent_trace_observability": 180,
+    "workflow-controls-v1/customer_service": 183,
+    "workflow-controls-v1/invoice_processing": 179,
+    "workflow-controls-v1/security_incidents": 174
+  },
+  "full_pass_four_gpu_steps": 37160,
+  "full_pass_four_gpu_rows_consumed": 148640,
+  "full_pass_four_gpu_wrapped_rows": 1
+}

provenance/wanli-manifest.json

@@ -0,0 +1,64 @@
+{
+  "schema_version": 1,
+  "configuration": {
+    "type": "import",
+    "version": "wanli-decisions-v1",
+    "seed": 20260921,
+    "repo": "alisawuffles/WANLI",
+    "revision": "61c95318fd71c55b6ba355d76253254615f387ec",
+    "license": "CC-BY-4.0",
+    "input_files_sha256": {
+      "train.jsonl": "85058cf017a911e89242dc29fa0a4ddaad3664cb923dc0a82145fdda14b694e5",
+      "test.jsonl": "4276e0af7fcdf657d1ab7beb54eaf025fda592a76c9ee86b63b7871953fc74fd",
+      "README.md": "63d95359bdd0c572c6196da6aff031593b1bc9d63c6d16d823a5f270b5ff0f9c"
+    },
+    "source_counts": {
+      "train": 102885,
+      "test": 5000
+    },
+    "removed": {
+      "conflicting_duplicate_pair": 2,
+      "official_test_seed_or_premise_component": 6675,
+      "duplicate_pair": 1
+    },
+    "limitations": [
+      "Human reviewed labels are not infallible.",
+      "One English NLI source; not an additional collection of operational domains.",
+      "No JevBench or frontier instances were supplied to this conversion."
+    ]
+  },
+  "summary": {
+    "records": 101207,
+    "groups": 57089,
+    "unique_inputs": 101207,
+    "splits": {
+      "calibration": 4707,
+      "ood": 4627,
+      "test": 4998,
+      "train": 82045,
+      "validation": 4830
+    },
+    "kinds": {
+      "choice": 101207
+    },
+    "families": {
+      "natural_language_inference": 101207
+    },
+    "sources": {
+      "wanli-decisions-v1": 101207
+    }
+  },
+  "files_sha256": {
+    "train.jsonl": "62b2561f1a43d157eb854916a4d24047f78e61dddb2e700b49d3268111449a85",
+    "calibration.jsonl": "7253def87adf070728e534f39284df1e349e30a13c1d65291d5189afc7f0aa70",
+    "validation.jsonl": "601ced227bb054b035654073c74eb4bafef29b3e368c325b083dab5c24fae384",
+    "test.jsonl": "5383a944978701468b8bdc884ed7ad41e78f2aea369317feb08b931af865de4f",
+    "ood.jsonl": "ce1fbebd84b6ca95d3dd94b2e93b224412a78090d9cafc2d4d560e61690aa13e"
+  },
+  "model_input_fields": [
+    "state",
+    "question",
+    "kind",
+    "options"
+  ]
+}

release-manifest.json

@@ -0,0 +1,90 @@
+{
+  "schema_version": 1,
+  "repo_id": "ZefanCai/Open-Jev-27B-v1.1",
+  "artifact_type": "qwen_lora_adapter_plus_scalar_decision_head",
+  "base_model": "Qwen/Qwen3.8-27B",
+  "base_revision": "1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
+  "base_weights_included": false,
+  "source_checkpoint_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+  "packaged_checkpoint_directory_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+  "checkpoint_tree_preserved": true,
+  "files": {
+    "LICENSE": {
+      "sha256": "bbedc3fda3305820b977265f01b8619d87570a6739de3a5582c3464840f1e57a",
+      "bytes": 11544
+    },
+    "LICENSE-CODE": {
+      "sha256": "ef2f6f6db605a2f611bc2d037c59276f615cdea76ac6e7551a28f71db890c9d2",
+      "bytes": 1078
+    },
+    "README.md": {
+      "sha256": "17a0f0474e9388b037f52efd7a5f11f66075a5fa60ff6cf44c88ec6b86c755af",
+      "bytes": 13542
+    },
+    "UPSTREAM.md": {
+      "sha256": "09a84a9466879daadb6eb0e74e14de9d76d6112345b47467fa5fedae1146119e",
+      "bytes": 1956
+    },
+    "package/checkpoint/adapter/README.md": {
+      "sha256": "f9e934433c6ae81516474d416c6b1dcf5193fe08bf41205363cf0b5b33760c1e",
+      "bytes": 5128
+    },
+    "package/checkpoint/adapter/adapter_config.json": {
+      "sha256": "0a11cc4a895a0ef163c1d25a4e2a5600c362adb805b4ce9e6278de9e8c78946a",
+      "bytes": 1170
+    },
+    "package/checkpoint/adapter/adapter_model.safetensors": {
+      "sha256": "1c857224bd3609c6a71eacf7f71dd021115fcc0f791936b1fc332e915b548a81",
+      "bytes": 61907944
+    },
+    "package/checkpoint/head.pt": {
+      "sha256": "76e382f122abfa4e0c467d860a8d142d2fb6d2a98dc0ef9e19870bfc6eb296b4",
+      "bytes": 22221
+    },
+    "package/checkpoint/model.json": {
+      "sha256": "394c4ccc4f880172571abd141e16d00bf9dc140d1e252bae4e9185ee4cf8c508",
+      "bytes": 187
+    },
+    "package/checkpoint/temperature.json": {
+      "sha256": "185c0b85539d195d02a2d4949295f0400fc208d9eb7ac4e5e6c97ebd339a235f",
+      "bytes": 162
+    },
+    "package/provenance.json": {
+      "sha256": "c69c18204588dcb0a3d42a9b0ba648561168de21cba2bdb7a6bf503d1a5c69a1",
+      "bytes": 15158
+    },
+    "provenance/lexical-overlap-screen.json": {
+      "sha256": "94fb13d6318a2afd536c594737e23f0227a9dc73d3bc739f4786e51aaf2e1332",
+      "bytes": 1128
+    },
+    "provenance/original-training-mixture-manifest.json": {
+      "sha256": "7b26f948d2ae11f20d7a18be437d1ada616fc479e1596ae94be7596787fe4e54",
+      "bytes": 24757
+    },
+    "provenance/wanli-manifest.json": {
+      "sha256": "38d50b03e62c63c561eff879351a73c426fec3a431b8cedd5edb0d564aa0f5af",
+      "bytes": 1962
+    },
+    "verification/checkpoint.json": {
+      "sha256": "360be115220ba4d98bcd486acb7cbd255de44f7ff15ca79e72501e9d51bc9eda",
+      "bytes": 1149
+    },
+    "verification/cpu.json": {
+      "sha256": "33a6fbb67e413fb0b8357f8d864e72d63de96ec007937ef4ee40bef4d68ab9d3",
+      "bytes": 909
+    },
+    "verification/internal-full.json": {
+      "sha256": "19a154c1a6ed3935bac266ab04b8a395bc180739540127b4b3802190e3c34f97",
+      "bytes": 4881
+    },
+    "verification/jevbench.json": {
+      "sha256": "9a78ffb3aecb113e26bf4d83fd5ad7467141e3bdf84535e90ba71d634ec2a85d",
+      "bytes": 6685
+    },
+    "verification/original-data.json": {
+      "sha256": "b067cb299ae1bd12b6740c1139ba432208777ede6db3066cc1334e39fa320cbe",
+      "bytes": 602
+    }
+  },
+  "scope": "All prepared payload files except release-manifest.json itself; no upload has occurred."
+}

verification/checkpoint.json

@@ -0,0 +1,18 @@
+{
+  "status": "passed",
+  "source_directory_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+  "packaged_directory_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+  "all_checkpoint_file_bytes_match_source": true,
+  "checkpoint_files_sha256": {
+    "adapter/README.md": "f9e934433c6ae81516474d416c6b1dcf5193fe08bf41205363cf0b5b33760c1e",
+    "adapter/adapter_config.json": "0a11cc4a895a0ef163c1d25a4e2a5600c362adb805b4ce9e6278de9e8c78946a",
+    "adapter/adapter_model.safetensors": "1c857224bd3609c6a71eacf7f71dd021115fcc0f791936b1fc332e915b548a81",
+    "head.pt": "76e382f122abfa4e0c467d860a8d142d2fb6d2a98dc0ef9e19870bfc6eb296b4",
+    "model.json": "394c4ccc4f880172571abd141e16d00bf9dc140d1e252bae4e9185ee4cf8c508",
+    "temperature.json": "185c0b85539d195d02a2d4949295f0400fc208d9eb7ac4e5e6c97ebd339a235f"
+  },
+  "checkpoint_tree_preserved": true,
+  "excluded_source_files": [],
+  "adapter_readme_review": "Original PEFT template inspected; no secrets or sensitive runtime paths found; preserved unchanged.",
+  "no_base_weights_optimizer_rng_credentials_or_raw_evaluation_records": true
+}

verification/cpu.json

@@ -0,0 +1,28 @@
+{
+  "status": "passed",
+  "method": "CPU safetensors and weights_only head load; pinned Qwen text architecture instantiated on meta",
+  "base_model_parameters_materialized": false,
+  "gpu_used": false,
+  "inference_rerun": false,
+  "adapter_tensors": 320,
+  "head_shapes": {
+    "weight": [
+      1,
+      5120
+    ],
+    "bias": [
+      1
+    ]
+  },
+  "trainable_parameters": 15471617,
+  "all_adapter_shapes_match_pinned_architecture": true,
+  "all_saved_tensors_finite_fp32": true,
+  "calibration_count_and_ids_verified": true,
+  "calibration_rows_belong_to_original_calibration_split": true,
+  "saved_temperature": 2.5343690298472983,
+  "training_completion_metadata_verified": true,
+  "limitations": [
+    "No fresh inference or numerical parity test on the 27B base model was run during packaging.",
+    "Full internal evaluation is a separate release gate and no incomplete metrics are included."
+  ]
+}

verification/internal-full.json

@@ -0,0 +1,127 @@
+{
+  "schema_version": 1,
+  "status": "passed_complete_full24_audit",
+  "publishable": true,
+  "model": "Open-Jev-27B-v1.1",
+  "checkpoint": {
+    "model": "Qwen/Qwen3.8-27B",
+    "base_revision": "1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
+    "checkpoint_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+    "temperature": 2.5343690298472983,
+    "lora_rank": 8,
+    "base_weights_included": false
+  },
+  "coverage": {
+    "total": 127787,
+    "successful": 127787,
+    "failed": 0,
+    "missing": 0,
+    "duplicates": 0,
+    "imported": 4491,
+    "new": 123296,
+    "completed_groups": 6,
+    "owned_processes_remaining": 0
+  },
+  "panels": {
+    "old_test": {
+      "count": 10532,
+      "hard_count": 10046,
+      "soft_count": 486,
+      "hard_correct": 9876,
+      "accuracy": 0.9830778419271352,
+      "expected_accuracy": 0.954681027079432,
+      "nll": 0.13386432263131773,
+      "brier": 0.03458594885426232,
+      "expected_brier": 0.06263691316614574,
+      "multiclass_ece": 0.029161848979411622,
+      "ece_bins": 15
+    },
+    "expanded_test": {
+      "count": 43301,
+      "hard_count": 42789,
+      "soft_count": 512,
+      "hard_correct": 41357,
+      "accuracy": 0.966533454859894,
+      "expected_accuracy": 0.9595090315974361,
+      "nll": 0.12983825446990038,
+      "brier": 0.05746167254927552,
+      "expected_brier": 0.06460776054107147,
+      "multiclass_ece": 0.037463708716989785,
+      "ece_bins": 15
+    },
+    "old_ood": {
+      "count": 15920,
+      "hard_count": 15446,
+      "soft_count": 474,
+      "hard_correct": 14825,
+      "accuracy": 0.9597954162890069,
+      "expected_accuracy": 0.9430203212128463,
+      "nll": 0.15887220783486702,
+      "brier": 0.06296576556098005,
+      "expected_brier": 0.08029650520798853,
+      "multiclass_ece": 0.024102283529718986,
+      "ece_bins": 15
+    },
+    "expanded_ood": {
+      "count": 84486,
+      "hard_count": 83924,
+      "soft_count": 562,
+      "hard_correct": 80934,
+      "accuracy": 0.9643725275249034,
+      "expected_accuracy": 0.9606666609107841,
+      "nll": 0.13434172410803066,
+      "brier": 0.060333553335607484,
+      "expected_brier": 0.06415555195289763,
+      "multiclass_ece": 0.01820277638023621,
+      "ece_bins": 15
+    }
+  },
+  "protocol": {
+    "max_length": 4096,
+    "truncation": false,
+    "cache": false,
+    "row_batch_size": 1,
+    "temperature_fitted_on_evaluation": false,
+    "base_dtype": "BF16",
+    "head_dtype": "FP32",
+    "hard_accuracy": "Exact one-hot targets only; all rows contribute to probability metrics.",
+    "subset": "Old Test/OOD are content-identical subsets of expanded Test/OOD; panels overlap."
+  },
+  "audit": {
+    "independent_audit_sha256": "650642bac1c7440f7cdbaf044068b8870ef7a514510dfd8a6cb01756a9304795",
+    "campaign_manifest_sha256": "491a4ee96d1ef6509ead733daffdf2318f3bb43d9d5bf38715f2cd0be8a0a9d4",
+    "capture_receipt_sha256": "bc90ec6386a1f2ec184ebc277df63c1016a3d0e32ac2a0661b3bd9371614d4a8",
+    "auditor_sha256": "c573ff8620e706197008170dc464e4e7a3de6da37fd93df58cc20ae020e8c393",
+    "arithmetic_sha256": "2eefd11d5010bbe705641f397eec0912992984ca3047631e8a8578eae5d34801",
+    "code_commit": "9eebd42de6137735fc5fd85b499aa1ffa256426d",
+    "merged_summary_sha256": "d657388b1142e226f81e3f3bc5b641a338d249f2ed936683bfdd3327b15bd20f",
+    "maximum_probability_error": 8.881784197001252e-16,
+    "maximum_metric_error": 0.0,
+    "metric_relative_absolute_tolerance": 3e-11,
+    "probability_absolute_tolerance": 1e-12,
+    "pairwise_probe_maximum_logit_error": 0.0,
+    "checks": [
+      "All raw IDs, order, assignment and unchanged bytes",
+      "Complete stable remote-before/after/local capture hashes",
+      "Original 4491 plus disjoint 123296 new predictions",
+      "All six successful controller exits and owned GPU cleanup",
+      "Frozen code, data and checkpoint identities and passing probes",
+      "Finite logits and saved-temperature softmax",
+      "Exact hard numerators and hard/soft denominators; full and per-source/type/question metrics"
+    ]
+  },
+  "data_scope": {
+    "original_mixture_rows": 328672,
+    "original_train_rows": 148639,
+    "public_projection_rows": 326619,
+    "public_train_rows": 147139,
+    "excluded_wiki_rows": 2053,
+    "note": "This internal evaluation uses the original complete mixture. The redistributable HF projection is not the complete training/evaluation corpus."
+  },
+  "limitations": [
+    "Shared multi-GPU timings do not establish single-GPU or HTTP/API latency or speedup.",
+    "New 2B training stopped; new 9B training not started. Released 2B/9B remain separate checkpoints.",
+    "JevBench is a separate public 231-task evaluation, not the complete 534-task benchmark.",
+    "Private raw/capture evidence is retained with hashes. The public aggregate and verifier code alone cannot reconstruct restricted source records."
+  ]
+}

verification/jevbench.json

@@ -0,0 +1,234 @@
+{
+  "schema_version": 1,
+  "status": "complete_independently_audited",
+  "model": "New Open-Jev 27B on community-hard-mix-v2",
+  "optimizer_steps": 37160,
+  "checkpoint": {
+    "model": "Qwen/Qwen3.8-27B",
+    "method": "lora_decision_head",
+    "checkpoint_sha256": "c49994563c3c4f04a99d9130203c4e526f4ae5086c84deec57698d18cb652e71",
+    "base_revision": "1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
+    "temperature": 2.5343690298472983,
+    "code_commit": "7ac6bab261bd97efc3d446152999e94ae432e5b0",
+    "max_length": 16384
+  },
+  "scope": "231 public JevBench tasks only: 72 original, 48 easy, 111 hard. The private and judge tiers are unavailable; this is not the full 534-task benchmark and no full-benchmark composite is computed. ZefanCai Open-Jev uses Qwen LoRA weights and a decision head; it is a different system from the Kotoba and Codiv projects also called OpenJev.",
+  "benchmark": {
+    "upstream_commit": "f8ce71361165846101d02ebc83ad44e47ae44fc3",
+    "input_sha256": "2e20ff5f94dd04d015b3a7b9abd955757b6c5be25b55b97fcd141cb52f88b4aa",
+    "dataset_sha256": "dc3995d8ae1e2fc8e81ce38431add509eb8bb39b85aadfd0c7c32079382dde51"
+  },
+  "metrics": {
+    "n_planned": 231,
+    "n_attempted": 231,
+    "n_scorable": 231,
+    "n_valid": 231,
+    "n_correct": 197,
+    "accuracy": 0.8528138528138528,
+    "schema_validity": 1.0,
+    "schema_validity_strict": 1.0,
+    "n_renormalized": 0,
+    "operational_success": 1.0,
+    "brier_mean": 0.2419831167860275,
+    "ordinal_mae": 0.3798003972534468
+  },
+  "per_public_tier": {
+    "original": {
+      "n_planned": 72,
+      "n_attempted": 72,
+      "n_valid": 72,
+      "n_correct": 69,
+      "accuracy": 0.9583333333333334
+    },
+    "easy": {
+      "n_planned": 48,
+      "n_attempted": 48,
+      "n_valid": 48,
+      "n_correct": 48,
+      "accuracy": 1.0
+    },
+    "hard": {
+      "n_planned": 111,
+      "n_attempted": 111,
+      "n_valid": 111,
+      "n_correct": 80,
+      "accuracy": 0.7207207207207207
+    }
+  },
+  "per_family": {
+    "policy": {
+      "n_planned": 12,
+      "n_correct": 11,
+      "accuracy": 0.9166666666666666
+    },
+    "intent": {
+      "n_planned": 24,
+      "n_correct": 23,
+      "accuracy": 0.9583333333333334
+    },
+    "ordinal": {
+      "n_planned": 12,
+      "n_correct": 12,
+      "accuracy": 1.0
+    },
+    "extraction": {
+      "n_planned": 24,
+      "n_correct": 23,
+      "accuracy": 0.9583333333333334
+    },
+    "adequacy": {
+      "n_planned": 12,
+      "n_correct": 12,
+      "accuracy": 1.0
+    },
+    "routing": {
+      "n_planned": 12,
+      "n_correct": 12,
+      "accuracy": 1.0
+    },
+    "fact": {
+      "n_planned": 12,
+      "n_correct": 12,
+      "accuracy": 1.0
+    },
+    "tool_selection": {
+      "n_planned": 12,
+      "n_correct": 12,
+      "accuracy": 1.0
+    },
+    "long_policy": {
+      "n_planned": 19,
+      "n_correct": 11,
+      "accuracy": 0.5789473684210527
+    },
+    "probability": {
+      "n_planned": 10,
+      "n_correct": 7,
+      "accuracy": 0.7
+    },
+    "temporal_numeric": {
+      "n_planned": 15,
+      "n_correct": 2,
+      "accuracy": 0.13333333333333333
+    },
+    "ambiguous": {
+      "n_planned": 7,
+      "n_correct": 5,
+      "accuracy": 0.7142857142857143
+    },
+    "multi_hop": {
+      "n_planned": 18,
+      "n_correct": 16,
+      "accuracy": 0.8888888888888888
+    },
+    "tradeoff": {
+      "n_planned": 6,
+      "n_correct": 5,
+      "accuracy": 0.8333333333333334
+    },
+    "adversarial": {
+      "n_planned": 6,
+      "n_correct": 6,
+      "accuracy": 1.0
+    },
+    "trap": {
+      "n_planned": 8,
+      "n_correct": 8,
+      "accuracy": 1.0
+    },
+    "judge_hard": {
+      "n_planned": 17,
+      "n_correct": 15,
+      "accuracy": 0.8823529411764706
+    },
+    "routing_hard": {
+      "n_planned": 5,
+      "n_correct": 5,
+      "accuracy": 1.0
+    }
+  },
+  "comparison_baselines": [
+    {
+      "label": "Released Open-Jev 2B",
+      "overall": {
+        "n_correct": 150,
+        "n_planned": 231,
+        "accuracy": 0.6493506493506493
+      },
+      "hard": {
+        "n_correct": 46,
+        "n_planned": 111,
+        "accuracy": 0.4144144144144144
+      }
+    },
+    {
+      "label": "Released Open-Jev 9B",
+      "overall": {
+        "n_correct": 179,
+        "n_planned": 231,
+        "accuracy": 0.7748917748917749
+      },
+      "hard": {
+        "n_correct": 66,
+        "n_planned": 111,
+        "accuracy": 0.5945945945945946
+      }
+    },
+    {
+      "label": "Jev 1.13.0",
+      "overall": {
+        "n_correct": 200,
+        "n_planned": 231,
+        "accuracy": 0.8658008658008658
+      },
+      "hard": {
+        "n_correct": 81,
+        "n_planned": 111,
+        "accuracy": 0.7297297297297297
+      }
+    }
+  ],
+  "protocol": {
+    "sum_tolerance_strict": 0.001,
+    "sum_tolerance_rounding": 0.02,
+    "argmax_tie_break": "lexicographically_smallest_label",
+    "score_accuracy": "argmax_level_equals_gold_level",
+    "probabilities": "native",
+    "concurrency": 1,
+    "retries": 0,
+    "warmups": 0,
+    "prefix_cache": false,
+    "latency_note": "Four-rank frozen-base FSDP2 execution on a shared node; one observation per task. Timing starts after dispatch and includes collective prediction, rank validation and response validation. No HTTP/network timing, warmups or retries. These diagnostics are not comparable to the old single-GPU HTTP latency, matched-hardware speedups, throughput or energy efficiency.",
+    "transport": "torch_distributed_collective",
+    "execution": {
+      "mode": "frozen_base_fsdp2",
+      "world_size": 4,
+      "backend": "nccl",
+      "node_shared_with_other_jobs": true,
+      "candidate_batch_size": 1,
+      "http_server_started": false
+    }
+  },
+  "execution": {
+    "mode": "frozen_base_fsdp2",
+    "world_size": 4,
+    "backend": "nccl",
+    "node_shared_with_other_jobs": true,
+    "candidate_batch_size": 1,
+    "http_server_started": false
+  },
+  "failed_requests": 0,
+  "pending_requests": 0,
+  "in_flight_requests": 0,
+  "audit": {
+    "file": "audit.json",
+    "sha256": "0759d6d9d457e0b823f823d8d9e7b7c9f0f4f3a9278f6cd32aac1e7b442392fc"
+  },
+  "source_summary_sha256": "ede7fda4e114c22bed02ec5a389a3bd4b9952eeeeed4a64e39cdce88d6589b8f",
+  "limitations": [
+    "231 public tasks only; unavailable private/judge tasks were not scored.",
+    "Four-rank shared-node direct-execution timings cannot be compared as speedups against the old single-GPU HTTP or hosted HTTPS timings.",
+    "Checkpoint hash verification establishes artifact identity; this offline audit does not rerun model inference or establish benchmark-training semantic separation.",
+    "Raw requests, labels, task IDs and responses remain in an ignored local run directory."
+  ]
+}

verification/original-data.json

@@ -0,0 +1,22 @@
+{
+  "train": {
+    "sha256": "959dac64adc1717d92e9bb80371d4b940b2b5994f9e36e469b3eaed85883fdff",
+    "rows": 148639
+  },
+  "calibration": {
+    "sha256": "6a02712bc6dce4c99cae3acfa54cf049553d12e4b795cfa4c170bd7de5732981",
+    "rows": 26764
+  },
+  "validation": {
+    "sha256": "743ec57c076fee300bfe3e3eb364a8fb629addfd185724c0b66d58633bcd9372",
+    "rows": 25482
+  },
+  "test": {
+    "sha256": "2cd250906060a570e0c596f7bb6ee7fb299fbf314e011a8c72ecb78d70248295",
+    "rows": 43301
+  },
+  "ood": {
+    "sha256": "b575698e0f1006713ca736b96d1d8eb43a54d052e266c0f363b027b288f54498",
+    "rows": 84486
+  }
+}