Duplicate from SHSLab/Open-Jev

.gitattributes

@@ -0,0 +1,60 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.avro filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text

LICENSE-CODE-MIT

@@ -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.

LICENSE-DATA

@@ -0,0 +1,3 @@
+Original generated Open-Jev records are dedicated under CC0 1.0 Universal.
+https://creativecommons.org/publicdomain/zero/1.0/legalcode
+This dedication does not relicense upstream wording, sources, game assets, model weights, or code. See README.md and THIRD_PARTY_NOTICES.md.

README.md

@@ -0,0 +1,308 @@
+---
+license: cc0-1.0
+language:
+  - en
+  - zh
+  - tr
+task_categories:
+  - text-classification
+tags:
+  - open-jev
+  - synthetic
+  - typed-decisions
+  - probability-estimation
+  - control
+size_categories:
+  - 100K<n<1M
+configs:
+  - config_name: release-v2-redistributable
+    default: true
+    data_files:
+      - split: train
+        path: data/release-v2-redistributable/train-*.parquet
+      - split: calibration
+        path: data/release-v2-redistributable/calibration-*.parquet
+      - split: validation
+        path: data/release-v2-redistributable/validation-*.parquet
+      - split: test
+        path: data/release-v2-redistributable/test-*.parquet
+      - split: ood
+        path: data/release-v2-redistributable/ood-*.parquet
+  - config_name: browser-drone-expansion-v1-redistributable
+    data_files:
+      - split: train
+        path: data/browser-drone-expansion-v1-redistributable/train-*.parquet
+      - split: calibration
+        path: data/browser-drone-expansion-v1-redistributable/calibration-*.parquet
+      - split: validation
+        path: data/browser-drone-expansion-v1-redistributable/validation-*.parquet
+      - split: test
+        path: data/browser-drone-expansion-v1-redistributable/test-*.parquet
+      - split: ood
+        path: data/browser-drone-expansion-v1-redistributable/ood-*.parquet
+  - config_name: citation-control-v1
+    data_files:
+      - split: train
+        path: data/citation-control-v1/train-*.parquet
+      - split: calibration
+        path: data/citation-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/citation-control-v1/validation-*.parquet
+      - split: test
+        path: data/citation-control-v1/test-*.parquet
+      - split: ood
+        path: data/citation-control-v1/ood-*.parquet
+  - config_name: entity-alignment-control-v1
+    data_files:
+      - split: train
+        path: data/entity-alignment-control-v1/train-*.parquet
+      - split: calibration
+        path: data/entity-alignment-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/entity-alignment-control-v1/validation-*.parquet
+      - split: test
+        path: data/entity-alignment-control-v1/test-*.parquet
+      - split: ood
+        path: data/entity-alignment-control-v1/ood-*.parquet
+  - config_name: amount-extraction-control-v1
+    data_files:
+      - split: train
+        path: data/amount-extraction-control-v1/train-*.parquet
+      - split: calibration
+        path: data/amount-extraction-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/amount-extraction-control-v1/validation-*.parquet
+      - split: test
+        path: data/amount-extraction-control-v1/test-*.parquet
+      - split: ood
+        path: data/amount-extraction-control-v1/ood-*.parquet
+  - config_name: email-selection-control-v1
+    data_files:
+      - split: train
+        path: data/email-selection-control-v1/train-*.parquet
+      - split: calibration
+        path: data/email-selection-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/email-selection-control-v1/validation-*.parquet
+      - split: test
+        path: data/email-selection-control-v1/test-*.parquet
+      - split: ood
+        path: data/email-selection-control-v1/ood-*.parquet
+  - config_name: phone-extraction-control-v1
+    data_files:
+      - split: train
+        path: data/phone-extraction-control-v1/train-*.parquet
+      - split: calibration
+        path: data/phone-extraction-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/phone-extraction-control-v1/validation-*.parquet
+      - split: test
+        path: data/phone-extraction-control-v1/test-*.parquet
+      - split: ood
+        path: data/phone-extraction-control-v1/ood-*.parquet
+  - config_name: context-retention-control-v1
+    data_files:
+      - split: train
+        path: data/context-retention-control-v1/train-*.parquet
+      - split: calibration
+        path: data/context-retention-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/context-retention-control-v1/validation-*.parquet
+      - split: test
+        path: data/context-retention-control-v1/test-*.parquet
+      - split: ood
+        path: data/context-retention-control-v1/ood-*.parquet
+  - config_name: sponsor-segment-control-v1
+    data_files:
+      - split: train
+        path: data/sponsor-segment-control-v1/train-*.parquet
+      - split: calibration
+        path: data/sponsor-segment-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/sponsor-segment-control-v1/validation-*.parquet
+      - split: test
+        path: data/sponsor-segment-control-v1/test-*.parquet
+      - split: ood
+        path: data/sponsor-segment-control-v1/ood-*.parquet
+  - config_name: silent-failure-control-v1
+    data_files:
+      - split: train
+        path: data/silent-failure-control-v1/train-*.parquet
+      - split: calibration
+        path: data/silent-failure-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/silent-failure-control-v1/validation-*.parquet
+      - split: test
+        path: data/silent-failure-control-v1/test-*.parquet
+      - split: ood
+        path: data/silent-failure-control-v1/ood-*.parquet
+  - config_name: ir-control-v1
+    data_files:
+      - split: train
+        path: data/ir-control-v1/train-*.parquet
+      - split: calibration
+        path: data/ir-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/ir-control-v1/validation-*.parquet
+      - split: test
+        path: data/ir-control-v1/test-*.parquet
+      - split: ood
+        path: data/ir-control-v1/ood-*.parquet
+  - config_name: mailroom-control-v1
+    data_files:
+      - split: train
+        path: data/mailroom-control-v1/train-*.parquet
+      - split: calibration
+        path: data/mailroom-control-v1/calibration-*.parquet
+      - split: validation
+        path: data/mailroom-control-v1/validation-*.parquet
+      - split: test
+        path: data/mailroom-control-v1/test-*.parquet
+      - split: ood
+        path: data/mailroom-control-v1/ood-*.parquet
+---
+
+# Open-Jev: typed decision datasets
+
+Open-Jev turns a state and a question into a typed decision: a yes/no probability, a distribution over choices, independent label probabilities, or a discrete numeric/ordinal decision. This repository publishes twelve separate, frozen data configs from the [Open-Jev project](https://github.com/Zefan-Cai/Open-Jev-Dev), together with original manifests, exact raw records, source code and reconstruction instructions.
+
+These are controlled, mostly synthetic tasks and reference labels. They are not official TypeSafe/Jev training data, model predictions, or evidence of general capability. Open-Jev is independently implemented and is not affiliated with TypeSafe.
+
+## Configs and exact split counts
+
+| Config | Train | Calibration | Validation | Test | OOD | Total |
+|---|---:|---:|---:|---:|---:|---:|
+| `release-v2-redistributable` | 79,116 | 4,672 | 3,723 | 10,356 | 15,701 | 113,568 |
+| `browser-drone-expansion-v1-redistributable` | 108,624 | 6,794 | 5,493 | 14,726 | 25,160 | 160,797 |
+| `citation-control-v1` | 2,520 | 200 | 180 | 300 | 800 | 4,000 |
+| `entity-alignment-control-v1` | 6,944 | 728 | 280 | 1,008 | 2,240 | 11,200 |
+| `amount-extraction-control-v1` | 32,984 | 2,232 | 992 | 3,472 | 9,920 | 49,600 |
+| `email-selection-control-v1` | 3,618 | 81 | 189 | 432 | 1,080 | 5,400 |
+| `phone-extraction-control-v1` | 12,350 | 855 | 380 | 1,615 | 3,800 | 19,000 |
+| `context-retention-control-v1` | 6,138 | 456 | 558 | 522 | 2,160 | 9,834 |
+| `sponsor-segment-control-v1` | 6,345 | 756 | 540 | 999 | 2,160 | 10,800 |
+| `silent-failure-control-v1` | 6,432 | 264 | 360 | 624 | 1,920 | 9,600 |
+| `ir-control-v1` | 7,366 | 464 | 464 | 986 | 2,320 | 11,600 |
+| `mailroom-control-v1` | 73,472 | 4,018 | 6,027 | 8,323 | 22,960 | 114,800 |
+
+**The configs overlap.** `release-v2-redistributable` is contained in `browser-drone-expansion-v1-redistributable`; do not add config totals and call them unique examples. Counts are typed decision rows. Several heads may come from the same conversation, document, family or game trajectory.
+
+The original frozen `release-v2` used to train the 2B/9B models has **80,816 training rows** and **115,821 rows across all splits**. The public projection above is not that exact training dataset. Each of the two composite configs excludes exactly **2,253 Wikispeedia rows**: 1,700 train, 89 calibration, 69 validation, 176 test and 219 OOD. The original expansion mixture has 163,050 rows, including 110,324 train. Original manifests and hashes are preserved without modification; [REPRODUCTION.md](REPRODUCTION.md) explains how to restore both exact original mixtures with separately obtained source data.
+
+The five citation/entity/amount/email/phone corpora were prepared and audited after those 2B/9B runs. **They have not been used for training or actual model inference at the time of this release.** The expansion mixture belongs to a separate 27B experiment; this dataset publication makes no completed-training or performance claim for that experiment.
+
+## Load and decode
+
+```python
+import json
+from datasets import load_dataset
+
+ds = load_dataset("ZefanCai/Open-Jev", "release-v2-redistributable")
+example = ds["train"][0]
+state = json.loads(example["state_json"])
+metadata = json.loads(example["metadata_json"])
+original_record = json.loads(example["record_json"])
+```
+
+For reproducibility, pass `revision="<dataset commit SHA>"`. Choose a config explicitly; loading the default does not include the separate control corpora.
+
+| Column | Meaning |
+|---|---|
+| `id`, `group_id`, `split`, `source` | Original identity, grouping, split and generator/source version. |
+| `kind` | Original decision type; interpret with the source task definition. |
+| `question`, `options` | Model-visible question and ordered answer space. |
+| `target` | Original numeric reference targets, represented as a float64 list. These are labels, not measured model confidence. |
+| `state_json` | JSON encoding of the original state. Decoding returns a string or structured object, depending on the source. |
+| `metadata_json` | Original provenance and audit metadata. It can include privileged teacher/control labels and must not be used as model input. |
+| `record_json` | Complete original JSON record, retaining original object key order and numeric representation. |
+| `original_line_number` | One-based row position in the original frozen split, including positions of excluded rows. |
+
+The Parquet representation avoids imposing one nested schema on different task states. `raw/<config>/<split>.jsonl.gz` preserves the source JSONL bytes after decompression. For the filtered composites, retained lines preserve exact bytes and order. For the ten stand-alone control configs, decompressed files match the original frozen split hashes exactly.
+
+Use only `state`, `question`, `kind` and `options` as model inputs. Do not expose `target`, `metadata`, identities, split assignments or provenance fields to the model. Distribution, binary, multilabel and ordinal targets have different semantics; do not reduce every row to a single-class accuracy calculation.
+
+## Domains and construction
+
+- The base release covers controlled customer-support routing and triage; local workflow decisions; geometric painting probability requests; Snake and tic-tac-toe; simplified T-Rex/runner and platformer controls; numeric ViZDoom Basic trajectories; and controlled reasoning decisions.
+- The expansion additionally includes controlled browser state/action and drone state/control examples. These represent the declared simulated task forms, not unrestricted browser use or real aircraft operation.
+- Citation data uses original policy documents, quotes, visible facts and claims, with supported/contradicted/insufficient decisions. It tests those controlled relations, not arbitrary factual verification.
+- Entity alignment uses original catalog families, records, aliases and visible matching policies with multiple decision heads.
+- Amount, email and phone extraction separate deterministic candidate generation from typed selection/attribute decisions. Known candidate misses and partial matches are retained rather than replaced using gold answers. Conditional attribute heads and omitted-supervision counts are documented in the original manifests.
+
+The new corpora include compressed documents/families/cases in `artifacts/`. These are reproduction/audit artifacts, not additional typed rows to add to the totals. Citation has 200 documents and 4,400 cases: 4,000 typed semantic cases plus 400 quote-not-found controls. Entity alignment has 200 families and 2,800 cases; amount has 200 families and 3,200 documents; email has 200 families and 2,800 documents; phone has 200 families and 4,000 documents.
+
+Original data and split policies are recorded per corpus in `provenance/original-manifests/`. Related documents/entities/trajectories remain grouped within splits. OOD is source-specific, commonly reserved wording, layouts, control families or goals, and is not a universal unseen-domain benchmark. Local export verification checks unique IDs and cross-split group separation within each config.
+
+## Evaluation boundaries
+
+Train, calibration, validation, test and OOD are published separately. Train on the train split; use calibration only for the declared calibration procedure and validation for model selection. Test/OOD labels are public, so future work must disclose any use of them for development. Scores measured on the original full frozen mixtures must not be described as scores on these smaller public projections without recomputation.
+
+No Jev Frontier 100 question/answer payload is included. The external [jev-frontier-100 benchmark](https://github.com/softpudding/jev-frontier-100) remains separate from training and generation. No official private examples, game ROMs, game assets, model weights or credentials are included.
+
+## Licensing and provenance
+
+Original generated records are marked **CC0-1.0** in their existing provenance. This dedication covers our generated content, not upstream wording, external assets, source data or model weights. Original source code is **MIT**. Customer-control provenance retains its original note that short upstream question descriptions come from TypeSafe documentation without a verified source license; this release does not relicense those descriptions. See [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md).
+
+The [Wikispeedia archive](https://snap.stanford.edu/data/wikispeedia.html) does not declare a verified separate graph/path redistribution license. Therefore its task rows are excluded from the two public mixture projections. We do not infer that a current Wikipedia license covers the archived graph/path dataset. We publish its original manifest, source URL, archive SHA-256, exact exclusion positions and a local reconstruction utility, not its graph, paths or task payload.
+
+Wikispeedia references:
+
+- West and Leskovec. *Human Wayfinding in Information Networks.* WWW 2012.
+- West, Pineau and Precup. *Wikispeedia: An Online Game for Inferring Semantic Distances between Concepts.* IJCAI 2009.
+
+`export-manifest.json` records original and public split counts/hashes, source counts, exclusion positions, source-code fingerprints and published file hashes. [REPRODUCTION.md](REPRODUCTION.md) documents exact restoration and generator commands. The dataset repository's Git commit pins this complete release.
+
+## Community task additions — 2026-09-20
+
+Three additional original CC0 control configs contribute **30,234 typed rows across 15 new splits**. That addition brought the repository to **10 configs and 50 splits**. Existing configs, payloads, manifests and the default remain unchanged.
+
+**These three additions have not been used for training or model evaluation.** Independent data audits validate their reference labels and split integrity; they are not model benchmarks. The released 2B/9B training mixture is unchanged.
+
+- [`context-retention-control-v1`](cards/context-retention-control-v1.md): Keep or discard eligible completed tool-call records and full outputs under an explicit fixed retention policy. The shared state contains context, goal and history; full tool outputs are omitted. Two Noul questions are built per eligible call. Labels follow visible goal dependency closure, exact-evidence needs and output recoverability. Pinned and pending calls are software gates, without model labels. Whole task graphs and their goal/recoverability counterfactuals share a split. OOD reserves diamond dependencies and wording. This finite synthetic task is not evidence of useful arbitrary-session compaction.
+- [`sponsor-segment-control-v1`](cards/sponsor-segment-control-v1.md): Categorize original timestamped transcript segments with one Choice per segment: sponsor, self_promo, intro, outro, recap, content or other. Paid third-party promotion needs affirmative funding evidence; creator-owned promotion is separate. A brand, discount code or promo marker alone does not establish sponsorship. Complete video families and payment-evidence counterfactuals share a split; OOD reserves complete sentence wording. The data are finite synthetic transcripts, not scraped videos or an audio/visual benchmark.
+- [`silent-failure-control-v1`](cards/silent-failure-control-v1.md): Ask one is_silent_failure Noul using only the exact response body string. The transport status is outside model input. Current maintenance, business rejection, a sign-in page replacing requested data, missing explicitly required receipts and unmet delivery requirements are contrasted with valid empty results, recovered history, quoted error text, accepted queued jobs and permitted partial results. Judge transport failures never become negative ground truth. Whole provider/contract families retain all counterfactuals; OOD reserves Chinese wording and different JSON/HTML/text layouts. This controlled grammar does not establish correctness for arbitrary APIs.
+
+The new original cases remain separate audit artifacts. Exact raw JSONL bytes, original manifests and an isolated source snapshot accompany the Parquet splits. [The additive export manifest](exports/community-data-20260920.json) records only these additions; the original `export-manifest.json` is retained byte for byte.
+
+
+## Original retrieval-control addition — 2026-09-20
+
+[`ir-control-v1`](cards/ir-control-v1.md) adds **11,600 typed rows across five
+splits**, representing 8,800 requests, 400 queries and 200 complete fictional
+system families. That addition brought the repository to **11 configs and 55 splits**.
+It supports pointwise Noul/Score, pairwise/setwise Choice and listwise
+Choice/Score task forms under an original four-level relevance rubric.
+Choice labels supervise best-passage selection, not full ranking or uncertainty.
+
+The full corpus passed an independent body-derived label audit. No training on
+these IR records is claimed. A separately frozen six-query Jev pilot is documented
+in the source project; it is not a full-corpus or TREC evaluation. No TREC data,
+qrels, provider responses or third-party examples are included. Frozen pilot
+test/OOD families remain held out in the expanded corpus.
+
+All existing dataset payloads, original manifests, config definitions, the default
+config and frozen model-training mixtures remain unchanged. The
+[IR export manifest](exports/ir-data-20260920.json) binds this addition only.
+
+
+## Original multilingual mailroom addition — 2026-09-20
+
+[`mailroom-control-v1`](cards/mailroom-control-v1.md) adds **114,800 typed rows
+across five splits**, from 11,600 original email requests and 400 complete
+English/Chinese/Turkish families. The repository now declares **12 configs and
+60 splits**. Each request retains the source-shaped two Choice and nine Noul
+questions. Nonbill/nonreceipt categories have no supervised label, and unchanged
+heads across taxonomy variants share their original row. There are 114,800
+unique ordered inputs, or 114,400 after also disregarding candidate order.
+
+The independent auditor derives labels from final visible email text and
+taxonomy. All frozen probe rows/cases keep their bytes and split assignments.
+No real mailbox contents, private source examples, attachments or provider
+responses are included. No training or full-corpus model evaluation is claimed;
+the separately documented 87-request Jev probe is not production mail-triage
+accuracy or end-to-end mailbox automation.
+
+All existing payloads, manifests, config definitions, the default config and
+frozen training mixtures remain unchanged, including the preceding IR addition.
+The [mailroom export manifest](exports/mailroom-data-20260920.json) binds only
+this new config and its auxiliary audit artifacts.

REPRODUCTION.md

@@ -0,0 +1,103 @@
+# Reproduction and original training-set restoration
+
+Download a pinned dataset revision before reproducing an experiment:
+
+```python
+from huggingface_hub import snapshot_download
+snapshot_download("ZefanCai/Open-Jev", repo_type="dataset",
+                  revision="<dataset commit SHA>", local_dir="open-jev-data")
+```
+
+`export-manifest.json` identifies every public artifact by SHA-256. Original frozen manifests are preserved unchanged in `provenance/original-manifests/`. Source code used during export is in `reproduce/source-code/` with file hashes in the export manifest; this snapshot is authoritative if the linked GitHub branch changes.
+
+## Exact published raw files
+
+Decompress `raw/<config>/<split>.jsonl.gz` to obtain native Open-Jev JSONL. For the five new configs the uncompressed SHA equals the original manifest SHA. For the two composite projections it equals `raw_uncompressed_sha256` in the export manifest, because only Wikispeedia rows are missing. Gzip encoding uses `mtime=0` and no embedded filename.
+
+Parquet retains the same records in `record_json`, and exposes structured top-level columns plus `state_json`/`metadata_json`. Parse those JSON columns before passing examples to the original code. Do not use audit metadata as model input.
+
+## Restore the exact original release-v2 and expansion mixtures
+
+The original `release-v2` includes 115,821 rows, and the original `browser-drone-expansion-v1` includes 163,050 rows. Both contain the same 2,253 Wikispeedia rows. Obtain those source data separately under their upstream terms:
+
+- Landing page: https://snap.stanford.edu/data/wikispeedia.html
+- Archive: https://snap.stanford.edu/data/wikispeedia/wikispeedia_paths-and-graph.tar.gz
+- Archive SHA-256: `97697096f5d2dcb77aa69e3992305c6c561de89edb9fb10b5ad9feaf8ba534d5`
+
+From `open-jev-data/reproduce/source-code`, with Python 3.10 or newer:
+
+```bash
+python -m jev.case_wikiracing --output-dir ../../../separately-obtained-wikiracing \
+  --targets 300 --pairs-per-target 12 --max-candidates 12 --seed 42
+```
+
+The generator fetches and verifies the fixed upstream archive. To use an existing archive, pass `--archive /path/to/wikispeedia_paths-and-graph.tar.gz` with the same SHA. This operation does not download any data from the public projections or modify a training directory.
+
+From the directory containing `open-jev-data` and `separately-obtained-wikiracing`:
+
+```bash
+python open-jev-data/reproduce/restore_original_mixture.py \
+  --release-root open-jev-data --config release-v2-redistributable \
+  --wiki-dir separately-obtained-wikiracing --output-dir restored-release-v2
+python open-jev-data/reproduce/restore_original_mixture.py \
+  --release-root open-jev-data --config browser-drone-expansion-v1-redistributable \
+  --wiki-dir separately-obtained-wikiracing --output-dir restored-browser-drone-expansion-v1
+```
+
+The utility checks every separate Wiki split against the original import manifest. It inserts Wiki rows at the recorded original positions, using exactly `jev.mix_data`'s JSON serialization, then verifies all ten restored split files against the original mixture hashes. It refuses to overwrite existing split files. This complete restoration was verified against the original frozen files during publication.
+
+Excluded row counts in each mixture are: train 1,700; calibration 89; validation 69; test 176; OOD 219. Their positions, but not their question/answer payload, are in `export-manifest.json`.
+
+## Original build commands
+
+The commands below document the original generator settings. Run in a separate workspace using the bundled source snapshot. Downloading/restoring the published frozen raw artifacts is the strongest byte-for-byte reproduction path; game generation additionally depends on its pinned runtime. The five newer manifests pin the relevant Python source hashes explicitly.
+
+Base mixture:
+
+```bash
+python -m jev.case_customer --output-dir data/case-customer --groups 1000
+python -m jev.case_workflows build --output-dir data/workflows-v1 --groups-per-workflow 250 --seed 42
+python -m jev.game_cli build-data all --output-dir data/games-v1 --episodes 100 --max-steps 80 --seed 42
+python -m jev.painting --output-dir data/painting-geometry-v1 --groups 60 --seed 42
+python -m jev.game_cli build-control-data --output-dir data/control-games-v1 --episodes 100 --max-steps 40 --seed 42
+python -m pip install vizdoom==1.2.4
+python -m jev.case_doom build --output-dir data/doom-basic-v1 --episodes 400 --ood-episodes 80 --seed 190919
+python -m jev.case_wikiracing --output-dir data/wikiracing --targets 300 --pairs-per-target 12 --max-candidates 12 --seed 42
+python -m jev.mix_data --inputs data/case-customer data/doom-basic-v1 data/wikiracing \
+  data/workflows-v1 data/painting-geometry-v1 data/games-v1 data/control-games-v1 --output-dir data/release-v1
+python -m jev.case_reasoning --output-dir data/reasoning-control-v1 --groups 2500 --seed 76109
+python -m jev.mix_data --inputs data/release-v1 data/reasoning-control-v1 --output-dir data/release-v2
+```
+
+Expansion:
+
+```bash
+python -m jev.case_browser --output-dir data/browser-v1 --groups 1000 --ood-groups 200 --seed 42
+python -m jev.case_drone --output-dir data/drone-control-v1 --groups 500 --ood-groups 100 --seed 42
+python -m jev.mix_data --inputs data/release-v2 data/browser-v1 data/drone-control-v1 \
+  --output-dir data/browser-drone-expansion-v1
+```
+
+Separately prepared new corpora:
+
+```bash
+python -m jev.case_citation --output-dir data/citation-control-v1 --groups 200 --ood-groups 40 --seed 42
+python -m jev.case_entity_alignment --output-dir data/entity-alignment-control-v1 --groups 200 --ood-groups 40 --seed 42
+python -m jev.case_amount_extraction --output-dir data/amount-extraction-control-v1 --groups 200 --ood-groups 40 --seed 42
+python -m jev.case_email_selection --output-dir data/email-selection-control-v1 --groups 200 --ood-groups 40 --seed 42
+python -m pip install phonenumbers==9.0.14
+python -m jev.case_phone_extraction --output-dir data/phone-extraction-control-v1 --groups 200 --ood-groups 40 --seed 42
+```
+
+This does not retrain any model. Check all generated files against the original manifest hashes before using them as a reproduction of a frozen corpus. Auxiliary case/document/family files in `artifacts/` can be decompressed directly; their uncompressed hashes are also recorded.
+
+## Rebuild this export from the frozen project workspace
+
+The GitHub repository keeps the export script and its input documentation in `reports/huggingface-data-release/`. With frozen source corpora available under that repository's `data/`, install `pyarrow`, then run:
+
+```bash
+python reports/huggingface-data-release/build_release.py \
+  --repo /path/to/Open-Jev-Dev --output /path/to/fresh-staging-directory
+```
+
+The exporter checks every frozen input split hash, applies the single documented source filter, validates all Parquet/raw round-trips and verifies both full-mixture restorations. Its allowlist excludes external benchmark payload, data drafts, game assets, credentials and model weights.

THIRD_PARTY_NOTICES.md

@@ -0,0 +1,14 @@
+# Third-party attribution and release boundaries
+
+Open-Jev is independently implemented and is not affiliated with TypeSafe.
+
+- TypeSafe Jev/System One documentation and the public Jev launch inspired the task forms. Links and pinned source revisions are in `docs/public-capabilities.md`. No proprietary RLCD code, weights or private training dataset is included. Publicly viewable evaluation examples have not been assigned a verified general redistribution/training license and are not copied into our training data.
+- `achimala/jev-paint` (previously `jevinci`), copyright Anshu Chimala, MIT, inspired the four pixel probability representations. Our request builders, simple mean-color renderer and geometry generator are independently written. We do not bundle its impasto renderer or recorded model-probability fixtures.
+- Community projects are attributed in `docs/games.md` and `docs/community.md`. External game code/ROMs/assets are not bundled. Our grid, runner and platformer engines are original simplified environments.
+- Qwen model weights and tokenizer files retain the terms of their exact source repositories. The three pinned revisions were each verified as Apache-2.0; [model provenance](docs/model-provenance.md) records the fixed sources, content hashes and attribution. Code licensing here does not relicense weights. Publication of adapters must include the applicable license text, modification notices and a model card referencing the corresponding base revision.
+- The exact common Qwen [Apache-2.0 license](third_party/qwen/LICENSE) and [attribution/modification notice](third_party/qwen/README.md) are included for checkpoint packaging. Inference weight bundles use Apache-2.0 and include the repository's MIT source-code license separately; generated-data CC0 declarations do not apply to weights.
+- ViZDoom and its bundled scenario assets retain their upstream per-file licenses. They are optional installed dependencies; see `docs/doom-case.md`.
+- The Wikispeedia graph and BoolQ auxiliary data retain their source licenses and attribution, recorded by import manifests. They are fetched by data scripts rather than copied into Git.
+- Generated geometry, controlled conversations, local game states and local workflow records are marked CC0-1.0 in their provenance. This covers our generated records, not an upstream source or model.
+
+The repository license applies to original code only. Check the manifests and source documentation before redistributing optional data, engines, checkpoints or assets.

artifacts/amount-extraction-control-v1/cases.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4bef25ef5110dd5a92a947f2818c577ec2c2ed9b4bc4cede6569b3be61e8e112
+size 1534317

artifacts/amount-extraction-control-v1/families.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70395bb3f90e967e3c4073925aecf00e81c02227e2d0dc5a78c87d874e7aa45e
+size 12879

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

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:218d10d30bad42e614b0de79184cda0c97720f43a42e6a382e6d1e0210c60a69
+size 178165

artifacts/citation-control-v1/documents.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68c9c964533c4da08201c0e112c1b694d073873c105918b891a2f65416fb974c
+size 45764

artifacts/context-retention-control-v1/cases.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e81c2639334e78f02a89fc3f01f723d3547aa433eb040145ecb07279d2d7ec0e
+size 515734

artifacts/email-selection-control-v1/cases.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df6d02e516c41f5cc8be25a3939246da693a6867817ecdea4c8b0018fcbf8f3c
+size 667160

artifacts/email-selection-control-v1/families.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c38d0896883d0d3d4f3213fca4eb833a07a17abee010f7e3b3099df7a15d7f7
+size 11756

artifacts/entity-alignment-control-v1/cases.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2be3eae5a179b9bf770b8b90c9c7a5487a1e7bca71f6bbeb44fc24cfc44be2d6
+size 178932

artifacts/entity-alignment-control-v1/families.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22f1a34e7c5a93369e25d530ca4b95a6f338b0a1d424a24957175ff78a2482e1
+size 40208

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

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

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

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

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

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7cda8e04e7bfc884d1ffba9e920c4108e45d80198e65952d3d1d4e93282f6b9
+size 2237258

artifacts/phone-extraction-control-v1/cases.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:81461836dbf03ac0d9de1b991e9508355e78755a56240bb93d13f42a27ec4ff2
+size 1166534

artifacts/phone-extraction-control-v1/families.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9dd99ceef0283f6905a32c506ba0b8b55af18c567f0b773cec57fefff7f9d0e7
+size 69065

artifacts/sponsor-segment-control-v1/cases.jsonl.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71772c2856104ce69fe052b1576ffb1efffda7504081f85e06b5acd4051c49a7
+size 663945

cards/context-retention-control-v1.md

@@ -0,0 +1,37 @@
+# context-retention-control-v1
+
+**9,834 original typed rows; 400 complete synthetic families.**
+Original generated content and labels are CC0-1.0. Source code keeps the repository's MIT license.
+
+**Prepared and independently audited only. Not used for training or model evaluation.**
+This addition changes neither the frozen release-v2 mixture nor the data used by the published 2B/9B checkpoints.
+
+Keep or discard eligible completed tool-call records and full outputs under an explicit fixed retention policy. The shared state contains context, goal and history; full tool outputs are omitted. Two Noul questions are built per eligible call. Labels follow visible goal dependency closure, exact-evidence needs and output recoverability. Pinned and pending calls are software gates, without model labels. Whole task graphs and their goal/recoverability counterfactuals share a split. OOD reserves diamond dependencies and wording. This finite synthetic task is not evidence of useful arbitrary-session compaction.
+
+| Split | Typed rows |
+| --- | ---: |
+| train | 6,138 |
+| calibration | 456 |
+| validation | 558 |
+| test | 522 |
+| ood | 2,160 |
+
+The community [interface reference](https://github.com/tamaratran/fast-jev-compaction/blob/e3f262a7f4d42bd8dd32ced30d26176f7cb545b0/src/compact.ts#L55) supplies the task shape.
+The task instances, labels and instructions are independently authored. No original third-party transcript, whole documentation page or model response is bundled.
+
+Use only state, question, kind and options as model inputs. Targets are reference labels; metadata and auxiliary cases may contain privileged information.
+The Parquet columns and JSON decoding rules match the root dataset card. Decompressing each raw split recovers its original JSONL bytes exactly.
+Whole-family splits include correlated counterfactuals; row counts must not be described as independently collected real-world cases.
+
+- [Original manifest](../provenance/original-manifests/context-retention-control-v1.json)
+- [Independent data audit](../provenance/community-data-20260920/context-retention-control-v1-audit.json)
+- [Exact generator source](../reproduce/community-data-20260920/source-code/jev/context_retention_data.py)
+- [Reproduction commands](../reproduce/community-data-20260920/source-code/README.md)
+
+From the source snapshot's root, regenerate in a new directory with:
+
+```bash
+python3 -m jev.context_retention_data --output-dir data/context-retention-control-v1 --groups 400 --ood-groups 80 --seed 942
+```
+
+No trained-model accuracy, useful compaction, natural-video classification, or general API failure-detection performance is claimed by this release.

cards/ir-control-v1.md

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

cards/mailroom-control-v1.md

@@ -0,0 +1,78 @@
+# Original multilingual mailroom controls
+
+**114,800 typed rows, 11,600 email requests and 400 whole synthetic families.**
+Original English, Chinese and Turkish emails and labels are CC0-1.0. Source
+code is MIT. No real mailbox, source emails, IMAP contents, attachments or the
+community author's private examples were imported.
+
+The [pinned community implementation](https://github.com/selcukusta/jev-mailroom/tree/06d44889231afb209e29275f14a529ba52c9eb0d)
+asks two Choice and nine Noul questions, eleven total. Our implementation
+preserves state shape `email.subject`, `email.from`, `email.date`, `email.body`
+and the eleven-head contract with independently authored question wording.
+
+`kind` distinguishes invoice, payment confirmation, account statement,
+promotion, newsletter and other. Four evidence Nouls ask about currently owed
+money, the sender issuing its own charges, billing identifiers and promotion.
+`category` distinguishes education, electricity, telecom, banking, airline and
+other; five independent Nouls also ask about these subjects. Category is
+unsupervised for nonbill/nonreceipt messages: 4,800 requests retain the runtime
+question but omit its training label. An application fallback is not ground truth.
+
+Already-paid amounts and advertised prices are not currently owed money.
+Invoice links without repeated amounts remain invoices; relaying a bill does
+not make the relay its issuer. Controls distinguish educational services from
+a student's phone bill, handset retail from telecom service, equipment from
+electricity supply, bank-mediated payment from a bank's own charges, and known
+out-of-taxonomy services. Unknown or contradictory grammar is not assigned gold.
+
+Each family contains three languages and nine actions, plus two English invoice
+taxonomy variants. Reordering categories tests order sensitivity; adding
+insurance changes the insurance label only when that option exists. No change
+in confidence is assumed or supervised. Shared unchanged question inputs across
+taxonomy variants reference their first row. There are **114,800 unique ordered
+inputs**, or **114,400** when candidate order is also disregarded.
+
+| Split | Typed rows |
+| --- | ---: |
+| train | 73,472 |
+| calibration | 4,018 |
+| validation | 6,027 |
+| test | 8,323 |
+| ood | 22,960 |
+
+All translations, actions, services and taxonomies for a family share a split.
+Index modulo ten values 8–9 reserve OOD; other families use stable hashing.
+OOD reserves fictional identities and moves body paragraphs when possible;
+two-paragraph messages retain their order. It is not a new grammar for every
+request. The frozen 10-family probe stays byte-identical in the expanded corpus,
+including all 290 requests and 2,870 rows in their original split files.
+
+Use only state, question, kind and options as model inputs. Targets, metadata,
+case reference labels and masks are privileged audit fields. Each Parquet
+`record_json` and decompressed raw split retain exact original JSONL bytes and
+row order, with the same column schema as the existing configs.
+
+**No training on this corpus is claimed.** A separate real Jev 1.13.0 probe
+completed 87 held-out requests containing 957 runtime questions and 921 labelled
+decisions; 36 category questions had no gold. Its 908/921 correct decisions are
+a small correlated control probe, not full-corpus evaluation or production
+multilingual accuracy. No provider responses are included in this data upload.
+The generation manifest records operations at generation time; later probe
+evaluation is documented separately. Frozen model-training mixtures are unchanged.
+
+- [Original manifest](../provenance/original-manifests/mailroom-control-v1.json)
+- [Independent body audit](../provenance/mailroom-data-20260920/independent-audit.json)
+- [Expansion preservation](../provenance/mailroom-data-20260920/expansion-stability.json)
+- [Auxiliary request cases](../artifacts/mailroom-control-v1/cases.jsonl.gz)
+- [Pinned generator](https://github.com/Zefan-Cai/Open-Jev/blob/61fc74408e95fa1844c1bd118245c2c871a84585/jev/mailroom_data.py), [auditor](https://github.com/Zefan-Cai/Open-Jev/blob/61fc74408e95fa1844c1bd118245c2c871a84585/jev/mailroom_audit.py) and [documentation](https://github.com/Zefan-Cai/Open-Jev/blob/61fc74408e95fa1844c1bd118245c2c871a84585/docs/mailroom-control-data.md)
+- [Incremental export manifest](../exports/mailroom-data-20260920.json)
+
+From the pinned source checkout, choose a fresh output directory:
+
+```bash
+python3 -m jev.mailroom_data --output-dir data/mailroom-control-v1-rebuilt --groups 400 --seed 943
+python3 -m jev.mailroom_audit --data data/mailroom-control-v1-rebuilt --output mailroom-audit.json
+```
+
+This is a finite original-data task, not a collection of natural private mail,
+live mailbox automation, or an independently annotated production benchmark.

cards/silent-failure-control-v1.md

@@ -0,0 +1,37 @@
+# silent-failure-control-v1
+
+**9,600 original typed rows; 400 complete synthetic families.**
+Original generated content and labels are CC0-1.0. Source code keeps the repository's MIT license.
+
+**Prepared and independently audited only. Not used for training or model evaluation.**
+This addition changes neither the frozen release-v2 mixture nor the data used by the published 2B/9B checkpoints.
+
+Ask one is_silent_failure Noul using only the exact response body string. The transport status is outside model input. Current maintenance, business rejection, a sign-in page replacing requested data, missing explicitly required receipts and unmet delivery requirements are contrasted with valid empty results, recovered history, quoted error text, accepted queued jobs and permitted partial results. Judge transport failures never become negative ground truth. Whole provider/contract families retain all counterfactuals; OOD reserves Chinese wording and different JSON/HTML/text layouts. This controlled grammar does not establish correctness for arbitrary APIs.
+
+| Split | Typed rows |
+| --- | ---: |
+| train | 6,432 |
+| calibration | 264 |
+| validation | 360 |
+| test | 624 |
+| ood | 1,920 |
+
+The community [interface reference](https://github.com/Vicente-MD/jev-resilience/blob/c490e0dc7830758f84bd9d5acb806655e327113e/src/main/java/ai/jev/resilience/client/JevEvaluationService.java#L50-L55) supplies the task shape.
+The task instances, labels and instructions are independently authored. No original third-party transcript, whole documentation page or model response is bundled.
+
+Use only state, question, kind and options as model inputs. Targets are reference labels; metadata and auxiliary cases may contain privileged information.
+The Parquet columns and JSON decoding rules match the root dataset card. Decompressing each raw split recovers its original JSONL bytes exactly.
+Whole-family splits include correlated counterfactuals; row counts must not be described as independently collected real-world cases.
+
+- [Original manifest](../provenance/original-manifests/silent-failure-control-v1.json)
+- [Independent data audit](../provenance/community-data-20260920/silent-failure-control-v1-audit.json)
+- [Exact generator source](../reproduce/community-data-20260920/source-code/jev/silent_failure_data.py)
+- [Reproduction commands](../reproduce/community-data-20260920/source-code/README.md)
+
+From the source snapshot's root, regenerate in a new directory with:
+
+```bash
+python3 -m jev.silent_failure_data --output-dir data/silent-failure-control-v1 --groups 400 --ood-groups 80 --seed 42
+```
+
+No trained-model accuracy, useful compaction, natural-video classification, or general API failure-detection performance is claimed by this release.

cards/sponsor-segment-control-v1.md

@@ -0,0 +1,37 @@
+# sponsor-segment-control-v1
+
+**10,800 original typed rows; 400 complete synthetic families.**
+Original generated content and labels are CC0-1.0. Source code keeps the repository's MIT license.
+
+**Prepared and independently audited only. Not used for training or model evaluation.**
+This addition changes neither the frozen release-v2 mixture nor the data used by the published 2B/9B checkpoints.
+
+Categorize original timestamped transcript segments with one Choice per segment: sponsor, self_promo, intro, outro, recap, content or other. Paid third-party promotion needs affirmative funding evidence; creator-owned promotion is separate. A brand, discount code or promo marker alone does not establish sponsorship. Complete video families and payment-evidence counterfactuals share a split; OOD reserves complete sentence wording. The data are finite synthetic transcripts, not scraped videos or an audio/visual benchmark.
+
+| Split | Typed rows |
+| --- | ---: |
+| train | 6,345 |
+| calibration | 756 |
+| validation | 540 |
+| test | 999 |
+| ood | 2,160 |
+
+The community [interface reference](https://github.com/valentynkit/jev-skip/blob/6837e3e0f1a48cbfc48c85415d99bcfe3eaf0628/lib/questions.ts) supplies the task shape.
+The task instances, labels and instructions are independently authored. No original third-party transcript, whole documentation page or model response is bundled.
+
+Use only state, question, kind and options as model inputs. Targets are reference labels; metadata and auxiliary cases may contain privileged information.
+The Parquet columns and JSON decoding rules match the root dataset card. Decompressing each raw split recovers its original JSONL bytes exactly.
+Whole-family splits include correlated counterfactuals; row counts must not be described as independently collected real-world cases.
+
+- [Original manifest](../provenance/original-manifests/sponsor-segment-control-v1.json)
+- [Independent data audit](../provenance/community-data-20260920/sponsor-segment-control-v1-audit.json)
+- [Exact generator source](../reproduce/community-data-20260920/source-code/jev/case_sponsor_segments.py)
+- [Reproduction commands](../reproduce/community-data-20260920/source-code/README.md)
+
+From the source snapshot's root, regenerate in a new directory with:
+
+```bash
+python3 -m jev.case_sponsor_segments --output-dir data/sponsor-segment-control-v1 --groups 400 --ood-groups 80 --seed 42
+```
+
+No trained-model accuracy, useful compaction, natural-video classification, or general API failure-detection performance is claimed by this release.

data/amount-extraction-control-v1/calibration-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73fa34f9a7c9a951efb542bfb91d360b40b8f926c121fef7b3dbe75e65d10937
+size 243408

data/amount-extraction-control-v1/ood-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ea2b28dc09d2048644f321cfecaaea149f025d28fdc5d91beef1ec76200f3a3
+size 1009686

data/amount-extraction-control-v1/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:deaa97a33988176b7c3022ab02b22b678ec8fd10f95d2010573097858fc9491f
+size 364630

data/amount-extraction-control-v1/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd09738e9463fd1efcdf87137600bea0c21270ee729c4475124d684b88d0cbb2
+size 3324405

data/amount-extraction-control-v1/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b555514e972eb0f38ac6e01691068aeb2f02b26e9e67754b13b908c80c4adc25
+size 100311

data/browser-drone-expansion-v1-redistributable/calibration-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20e24d12d308841b5d3249e22a16832625436e76861bf39742d436bcb4e1c557
+size 711600

data/browser-drone-expansion-v1-redistributable/ood-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f03b875a21e8a4c1868c8334455fe78d97b6a6a74f8a12b87374fa1d01f6d71
+size 2539241

data/browser-drone-expansion-v1-redistributable/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afc858fc1724a085375333d14bae5e4750fea4ab934f316e9b9b4c9890921a89
+size 1451091

data/browser-drone-expansion-v1-redistributable/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4fdfdff0d9f603d3e7ba51d5a48b2cd0d8eeb60f3244779b6b04d135f4a89f7
+size 9632515

data/browser-drone-expansion-v1-redistributable/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a70ba9b5700c128e529baf5cca6224743820d5501fe246b6d898851651f141ef
+size 594976

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

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f420cc1b004f6ba2024e96d5c64451a3434acd5759be8d0718e28b9fcd10ed2
+size 43744

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

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59c526a897b449d45483d7c26c5e8e55e609175e7e6c3b5e2375346da420c183
+size 91356

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

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e77adf6898d73beed303edd6bcceb9e1cd6fb6a01b9055c3804dc46acd225c91
+size 52200

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

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8402d2668c4c46bf9f5c3a8bc292652c8de72d288e8e5c0df30fd13d73b5c51d
+size 291394

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

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:faa557577101f587d6bd623136c32181c0474077e34c1168a400d458a1719f80
+size 42054

data/context-retention-control-v1/calibration-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:852a87c4816242f656ace8f7ab31c7d39e3643da452ec6bb0f3d40d1da8bbf3a
+size 59259

data/context-retention-control-v1/ood-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50f3eae83f6b02f5c53ad80693bff2c1959255449567e671ab93da30319750a5
+size 243289

data/context-retention-control-v1/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c65123feabbabe52abf5573e6345a7922d42c5c68cdfd98bd7d3476f212434c
+size 80850

data/context-retention-control-v1/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad416a9cb0bd7692ef978c799458edb0fa6c4fbc7b545355c3ccd3a52e420180
+size 719330

data/context-retention-control-v1/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e5fb2ce3d231856a1cd1fa74f0b8ced4b76cec9ec6d0b43f7450047ab62ef57
+size 83436

data/email-selection-control-v1/calibration-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fca99e888109d3a4e223e6b38f562c1c0f2563d62a48b831d778d10d10b8ac3
+size 41177

data/email-selection-control-v1/ood-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb74db37451174ce8684eeddf0e846b80d65e658e2b74c3697a77aa8d42e948b
+size 256046

data/email-selection-control-v1/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b2aff2ca90e68d8346b4ea5cc23b045aad3f786a27371f8d9edb55feb677e02
+size 109919

data/email-selection-control-v1/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c91b6293515a21e709dd8fe5a05a05374e20933f0f8f6c4c847ec96707188c0
+size 809247