Publish data-only Jev Decisions v1 dataset

README.md

@@ -0,0 +1,251 @@
+---
+pretty_name: Jev Decisions v1
+license: cc-by-4.0
+license_link: https://creativecommons.org/licenses/by/4.0/
+task_categories:
+  - reinforcement-learning
+tags:
+  - agents
+  - tool-use
+  - function-calling
+  - software-engineering
+  - routing
+  - decision-making
+  - value-modeling
+  - parquet
+configs:
+  - config_name: default
+    data_files:
+      - split: train
+        path: data/train/*.parquet
+      - split: validation
+        path: data/validation/*.parquet
+      - split: test
+        path: data/test/*.parquet
+---
+
+# Jev Decisions v1
+
+**12M canonical agent-decision records for tool selection, routing, value prediction, completion, and local agent control.**
+
+Jev Decisions v1 is a derived, decision-oriented corpus built from public agent trajectory datasets. It canonicalizes heterogeneous trajectories into a shared learning interface:
+
+`state + available candidate decisions -> target / outcome / eligibility`
+
+The purpose is to make agent-control training easier without requiring each researcher to independently acquire and normalize several large, structurally different sources. The corpus is designed for action selectors, tool and function routers, model routers, next-action predictors, completion policies, value models, verification systems, and small local decision models. It is not primarily a conversational language-model pretraining corpus, and ordinary assistant prose is not the target.
+
+## Scale and supervision
+
+| Measure | Count |
+| --- | ---: |
+| Usable canonical records | 11,978,080 |
+| Choice-supervised decisions | 6,214,185 |
+| Value-supervised records | 9,369,747 |
+| Completion-supervised records | 48,293 |
+| Score-supervised records | 0 |
+| Unique trajectories/tasks | 1,066,352 |
+| Original canonical Parquet artifact | 48,763,693,033 bytes (45.415 GiB) |
+
+The deterministic group-aware split is approximately 94/3/3:
+
+| Repository split | `training_split` value | Records |
+| --- | --- | ---: |
+| train | `train` | 11,269,408 |
+| validation | `val` | 354,037 |
+| test | `test` | 354,635 |
+
+All rows from a trajectory/task stay in one partition. Verified trajectory/task overlap across train, validation, and test is zero. The repository uses the conventional `validation/` directory; the preserved field value is `val`.
+
+## Source composition
+
+The corpus currently derives from four NVIDIA-developed upstream dataset repositories:
+
+* [nvidia/Nemotron-SFT-Agentic-v2](https://huggingface.co/datasets/nvidia/Nemotron-SFT-Agentic-v2), configs/splits `default/interactive_agent`, `default/search`, and `default/tool_calling`.
+* [nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1](https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1), `default/train`.
+* [nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1](https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1), `default/train`.
+* [nvidia/Open-SWE-Traces](https://huggingface.co/datasets/nvidia/Open-SWE-Traces), version `v1.0`, `openhands` and `sweagent`.
+
+The card does not declare an English-only language tag. Open-SWE records include a repository programming-language field, and this corpus combines natural-language agent states with code and tool schemas; the description language should not be read as a claim that every source example is English-only.
+
+Choice supervision is composed of:
+
+| Source | Choice-eligible rows | Share |
+| --- | ---: | ---: |
+| Open-SWE-Traces | 3,605,852 | 58.03% |
+| Nemotron-SFT-Agentic-v2 | 2,537,900 | 40.84% (~41%) |
+| Nemotron Pivot datasets, combined | 70,433 | 1.13% (~1%) |
+
+Open-SWE supplies all current Value supervision. As a result, uniform multi-objective training is heavily weighted toward software engineering.
+
+## Dataset balance warning
+
+> **The complete dataset intentionally preserves the natural source distribution and is NOT source-balanced.** Open-SWE represents 78.22% of usable records, 58.03% of Choice supervision, and 100% of Value supervision. For general-purpose tool selection or routing, source-balanced or domain-balanced sampling may be preferable. Keep source and domain provenance when sampling.
+
+## Schema
+
+The Parquet schema is preserved in every shard. JSON-shaped payloads whose actual schema type is `string` remain strings; parse those fields explicitly if structured objects are needed. The schema does not contain one generic `source_config` field at the top level: source config and raw split are under `provenance`.
+
+| Field | Actual Parquet type / location | Meaning |
+| --- | --- | --- |
+| `id` | string | Canonical record ID. |
+| `source` | string | Upstream dataset ID. |
+| `decision_type` | string | Record kind, including `tool_choice`, `completion`, and `unsupported`. |
+| `status` | string | Canonical record status. |
+| `unsupported_reason` | nullable string | Reason an otherwise preserved record is unsupported. |
+| `content_hash` | string | Canonical content fingerprint. |
+| `state.system` | nullable string | System instructions available before the decision. |
+| `state.user_goal` | nullable string | User/task request available before the decision. |
+| `state.history` | list of structs `{role: nullable string, payload_json: string}` | Prior conversation/environment history. Each payload is serialized JSON text. |
+| `state.environment_json` | nullable string | Serialized environment state available before the decision. |
+| `candidates` | list of structs `{id: string, name: string, description: nullable string, parameters_json: string, metadata_json: string}` | Actions actually available at the decision. Parameter and metadata payloads are JSON strings. |
+| `target` | nullable struct `{candidate_id: nullable string, action_name: string, arguments_json: nullable string, label_json: nullable string}` | Single target action and its arguments/label where represented. |
+| `ordered_targets` | list of structs with the same four fields as `target` | Source-ordered target(s); use the single-choice eligibility flag before treating these as atomic Choice targets. |
+| `labels.complete` | nullable boolean | Completion label when supported by source evidence. |
+| `labels.trajectory_success` | nullable boolean | Verified trajectory outcome; null means unknown. |
+| `labels.value_target`, `labels.reward`, `labels.score`, `labels.source_pass_rate` | nullable float64 | Value/reward/score and source-native pass rate when present. These quantities are not interchangeable. |
+| `labels.outcome_evidence` | nullable string | Evidence supporting outcome labels. |
+| `provenance.dataset_id` | string | Exact upstream dataset ID. |
+| `provenance.dataset_revision` | string | Source revision/version recorded by canonicalization. |
+| `provenance.source_config`, `provenance.raw_split` | string | Upstream configuration and split. |
+| `provenance.source_row_index` | int64 | Row index in the source shard/split. |
+| `provenance.source_identity`, `provenance.trajectory_id` | string | Original source/task identity and trajectory grouping key. |
+| `provenance.step_index`, `provenance.decision_ordinal` | int64 | Step and canonical decision order. |
+| `provenance.adapter_version` | string | Adapter version recorded during canonicalization. |
+| `provenance.license_metadata_json`, `provenance.source_metadata_json` | string | Serialized source license and other source metadata. |
+| `training.choice_eligible` | boolean | Whether the row has supported Choice supervision. |
+| `training.completion_eligible` | boolean | Whether completion supervision is supported. |
+| `training.value_eligible` | boolean | Whether value supervision is supported. |
+| `training.score_eligible` | boolean | Whether score supervision is supported. |
+| `training.boolean_eligible`, `training.arguments_eligible` | boolean | Eligibility for those auxiliary targets. |
+| `training.use_for_bc`, `training.use_for_value` | boolean | Canonical training-use flags. |
+| `training.supervision_evidence` | string | Evidence category for the available training target. |
+| `training.choice_ineligible_reason` | nullable string | Reason Choice supervision is ineligible. |
+| `training.quality_weight` | float64 | Canonical record quality weight. |
+| `training_split` | nullable string | Existing deterministic assignment: `train`, `val`, or `test`. |
+
+Eligibility flags are authoritative: do not treat every row or every target as positive supervision. In particular, do not train Choice on failed or unresolved trajectory actions. A failed trajectory alone does not prove each preceding action was wrong.
+
+## Usage
+
+### Hugging Face Datasets
+
+Loading the whole dataset materializes a large dataset. For selective access without loading every row into RAM, use streaming below.
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("samatv256/jev-decisions-v1")
+train = ds["train"]
+
+choice = train.filter(
+    lambda row: row["training"]["choice_eligible"]
+)
+```
+
+### Streaming Choice-eligible rows
+
+```python
+from datasets import load_dataset
+
+train = load_dataset(
+    "samatv256/jev-decisions-v1",
+    split="train",
+    streaming=True,
+)
+
+for row in train:
+    if row["training"]["choice_eligible"]:
+        # Consume one row at a time; do not accumulate the entire corpus.
+        use_for_choice_training(row)
+```
+
+`training_split` is retained inside each row and can also be checked while streaming. The loader's `train` split maps to files under `data/train/`.
+
+### PyArrow
+
+```python
+import pyarrow.dataset as pads
+
+dataset = pads.dataset(
+    "hf://datasets/samatv256/jev-decisions-v1/data/train",
+    format="parquet",
+)
+eligible = dataset.scanner(
+    filter=pads.field("training", "choice_eligible") == True,
+    batch_size=1024,
+).to_batches()
+for batch in eligible:
+    consume(batch)
+```
+
+For broad compatibility across PyArrow versions, filtering each scanned batch is also valid:
+
+```python
+for batch in dataset.scanner(batch_size=1024).to_batches():
+    mask = batch.column("training").field("choice_eligible")
+    consume(batch.filter(mask))
+```
+
+### Polars
+
+```python
+import polars as pl
+
+lf = pl.scan_parquet(
+    "data/train/*.parquet",
+    hive_partitioning=False,
+)
+choice = lf.filter(pl.col("training").struct.field("choice_eligible"))
+for batch in choice.collect_batches(chunk_size=1024):
+    consume(batch)
+```
+
+For the hosted repository, use the Hub parquet URLs returned by the Hub API, or download the selected split first. `scan_parquet` is lazy; avoid calling `.collect()` on all rows unless the available memory is sufficient.
+
+## Training guidance
+
+### Choice, tool selection, and routing
+
+Filter on `training.choice_eligible == true`. The conceptual input is the state plus the available candidate set; the target is the correct candidate/action in `target` (or the source-ordered target representation where the row is eligible). Candidate names, descriptions, and parameter schemas are part of the decision problem. This supervision supports tool selection, function/API routing, and next-action prediction. Model routing can use the same form after adding candidate models and model-performance labels.
+
+### Value
+
+Filter on `training.value_eligible == true`. All current value-eligible records come from Open-SWE. The numeric target is source-supported trajectory outcome supervision, not a generally calibrated probability across all domains.
+
+### Completion
+
+Filter on `training.completion_eligible == true` for completion control. In v1 these are success-backed terminal `finish` examples; the corpus has no currently eligible verified-failure `continue` examples, so this head is positive-only and should not be treated as a balanced continue/finish classifier. Do not infer completion from an agent merely stopping. Do not treat unknown-outcome or failed-trajectory actions as positive Choice labels.
+
+## Integrity and exclusions
+
+* Original canonical artifact SHA256: `c4c19101d0a7e5a0e57428704214b73eed8fb559cdb4db70293b16bec1231593`
+* Raw canonical records processed: 15,251,846
+* Usable records: 11,978,080
+* Excluded/preserved ineligible records: 3,273,766
+* Exact duplicate IDs: 0
+* Conflicting duplicate IDs: 0
+* Train/validation/test trajectory overlap: 0
+
+Excluded or preserved ineligible examples include unknown/unresolved trajectory outcomes, reasoning-only turns, multi-action records unsupported by the current atomic Choice target, non-generative targets, and targets missing from candidate sets. The build preserved these as ineligible records in its reconciliation; this public v1 contains the usable canonical artifact only. No failed or unresolved trajectory actions were silently promoted to positive Choice supervision.
+
+The shards are a deterministic re-encoding of the original Parquet rows, partitioned only by the existing `training_split`. They preserve all columns and nested schema, every record exactly once, and row order within each split. There is no resampling or rebalancing. Most full training shards are about 1.3–2.8 GB; source record-size variation creates a few larger-than-target shards, and the final 19,408-row remainder is kept as a small shard rather than inflating its neighbor. Validation and test are each one shard. See [`publication_manifest.json`](publication_manifest.json) for per-shard row counts, byte sizes, split assignments, and SHA256 values. The manifest also contains matching ordered record-ID hashes per split; canonical IDs were verified unique before publication, so this checks that no record was omitted, duplicated, or reordered. During the build, each full shard was also re-read with PyArrow to verify its file hash, schema, row count, and split purity. The writer routes unmodified Arrow table slices, preserving all remaining field values.
+
+The canonicalizer constructs each visible state from the ordered history prefix ending before its target decision. Current/future tool actions, tool results, evaluator outcomes, reference patches, and resolution fields are kept in targets, labels, or provenance. It mechanically scans state for forbidden evaluator/reference keys, and publication validation confirms that sharding did not alter row content or order. `training_split` is a separate top-level field, not part of `state`; consumers should select a split before model input construction and must not feed labels or provenance into the model.
+
+## Attribution and license
+
+This is a derived transformation/canonicalization corpus. NVIDIA is the developer of the four upstream datasets listed above, not of Jev Decisions v1. This repository is independently prepared and published by `samatv256`; it is not endorsed by or affiliated with NVIDIA.
+
+At publication time, all four upstream Hugging Face dataset cards identify CC BY 4.0 as their dataset license. Their cards additionally identify Apache 2.0 and/or MIT licensing information; Open-SWE-Traces also carries MIT, Apache 2.0, BSD-2-Clause, and BSD-3-Clause additional licensing information and a per-repository SPDX `license` field. These terms are retained as provenance and are not replaced by this corpus's dataset-level attribution. In particular, the Open-SWE `provenance.source_metadata_json` / `license_metadata_json` fields preserve source metadata, including the repository license. Users must retain CC BY attribution and meet any applicable component, repository, and upstream terms when using or redistributing source-derived content. See [`SOURCE_LICENSES.md`](SOURCE_LICENSES.md).
+
+This dataset card does not grant rights beyond the upstream terms. NVIDIA's inclusion as a source developer does not imply endorsement or affiliation.
+
+## Repository contents
+
+* `data/train/`, `data/validation/`, `data/test/`: publication-friendly Parquet shards.
+* `publication_manifest.json`: machine-readable shard inventory, checksums, schema and row-content reconciliation.
+* `input_shards_manifest.json`: source shard inventory from the frozen canonical build.
+* `audit_sample_100.md`: deterministic 100-row human-readable audit sample; long fields are clipped for display only.
+* `source_revisions.json` and `SOURCE_LICENSES.md`: canonical source pins and pre-publication license metadata.
+
+This repository contains the dataset, its data card, source and license metadata, integrity manifests, and a small audit sample. It does not publish model results, experimental reports, or retraining materials.

SOURCE_LICENSES.md

@@ -0,0 +1,14 @@
+# Source licenses and attribution
+
+License metadata was re-checked on 2026-09-22 immediately before preparing the public release using each current Hugging Face dataset card. All four cards identify Creative Commons Attribution 4.0 International (CC BY 4.0) for the dataset. Some cards also list additional open-source licenses. Open-SWE-Traces preserves a per-repository SPDX license field in its source data; this per-repository information is not collapsed into a new blanket license.
+
+| Upstream dataset | Dataset card license | Additional licensing information on card | Attribution |
+| --- | --- | --- | --- |
+| [nvidia/Nemotron-SFT-Agentic-v2](https://huggingface.co/datasets/nvidia/Nemotron-SFT-Agentic-v2) | CC BY 4.0 | Apache 2.0; MIT | NVIDIA is upstream data developer. |
+| [nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1](https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1) | CC BY 4.0 | None additionally listed in the current card's terms section. | NVIDIA is upstream data developer. |
+| [nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1](https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1) | CC BY 4.0 | Apache 2.0; MIT | NVIDIA is upstream data developer. |
+| [nvidia/Open-SWE-Traces](https://huggingface.co/datasets/nvidia/Open-SWE-Traces) | CC BY 4.0 | MIT; Apache 2.0; BSD 2-Clause; BSD 3-Clause. Source rows also contain an SPDX repository license. | NVIDIA is upstream data developer. |
+
+Jev Decisions v1 is a derived canonicalization and selection corpus, not an NVIDIA dataset. NVIDIA is not represented as the author, publisher, or endorser of this derived work. This release preserves source dataset IDs, revisions, configs/splits, and license metadata in Parquet provenance fields and provides the original frozen source-shard manifest.
+
+The Hugging Face `license: cc-by-4.0` card field records the common upstream dataset-card license. It does not remove, replace, or override any additional applicable license or source-specific terms. Users should inspect row-level Open-SWE repository license metadata and comply with attribution and all applicable upstream terms.

data/test/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28977886b75a6e6c6fc861ca34f959a7de9e9b56047f77c2569cd175521121c5
+size 569190836

data/train/train-00000-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99424862fa54a112305df3977e19e3b8d7c7cdd9d310917ebace49c824d60364
+size 2015436866

data/train/train-00001-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7681974168ba0be686eabb452d093d7fae84b55320aa2c5891f1a3744cb5f14
+size 1340995814

data/train/train-00002-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:537c60828eb9c4c54b6c5a36e26a92245d098cccf0bee04aa047527e94d1f6b7
+size 1912265369

data/train/train-00003-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab34af18330635ba697920ea2a04c82a1a8d1bf5f810e87776b90754ac6f191d
+size 2078724448

data/train/train-00004-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28e8ffdd578799568b68b3ab946017aa24f2226cfea751f30566ec94a8f16c48
+size 2135350586

data/train/train-00005-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8983112410bedd4eab6f05ffb206b95531890324b2225661d4ee8add2ba2e700
+size 2155770487

data/train/train-00006-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b163351f7a6b43e524e3af717fd6a007aaac629f804cf3f42447bf4dd8d10323
+size 1996690317

data/train/train-00007-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0c62fdcb51da61ba663200230641d47d1f9fa08444336ac45124be0b16696ba
+size 1896341693

data/train/train-00008-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:203693c7c0b8642c4181fc7a9e90630c8d1db308eb4ad2f91c15c7c321609f2c
+size 2529808650

data/train/train-00009-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f31fcbfff51c311d81bb110dc0f253b70fd2992f08300101452415f5ef7e267
+size 2775383707

data/train/train-00010-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2011b96b14e417c351a6e86d5ff2b22802d2ef697fb2cf073b9f0550d624f6a8
+size 42750853

data/validation/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d9b37d703f51c0cc31ad788b92b656417181f399946ee4f951e9bc6df204ef4
+size 570669729

publication_manifest.json

@@ -0,0 +1,201 @@
+{
+  "dataset_id": "samatv256/jev-decisions-v1",
+  "artifact": "ips_all_ready_v1.parquet",
+  "original_artifact_bytes": 48763693033,
+  "original_artifact_sha256": "c4c19101d0a7e5a0e57428704214b73eed8fb559cdb4db70293b16bec1231593",
+  "records": 11978080,
+  "split_column": "training_split",
+  "split_counts": {
+    "train": 11269408,
+    "validation": 354037,
+    "test": 354635
+  },
+  "shard_target_rows": 1125000,
+  "shards": [
+    {
+      "filename": "data/test/test-00000-of-00001.parquet",
+      "rows": 354635,
+      "bytes": 569190836,
+      "sha256": "28977886b75a6e6c6fc861ca34f959a7de9e9b56047f77c2569cd175521121c5",
+      "split": "test"
+    },
+    {
+      "filename": "data/train/train-00000-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 2015436866,
+      "sha256": "99424862fa54a112305df3977e19e3b8d7c7cdd9d310917ebace49c824d60364",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00001-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 1340995814,
+      "sha256": "f7681974168ba0be686eabb452d093d7fae84b55320aa2c5891f1a3744cb5f14",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00002-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 1912265369,
+      "sha256": "537c60828eb9c4c54b6c5a36e26a92245d098cccf0bee04aa047527e94d1f6b7",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00003-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 2078724448,
+      "sha256": "ab34af18330635ba697920ea2a04c82a1a8d1bf5f810e87776b90754ac6f191d",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00004-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 2135350586,
+      "sha256": "28e8ffdd578799568b68b3ab946017aa24f2226cfea751f30566ec94a8f16c48",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00005-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 2155770487,
+      "sha256": "8983112410bedd4eab6f05ffb206b95531890324b2225661d4ee8add2ba2e700",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00006-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 1996690317,
+      "sha256": "b163351f7a6b43e524e3af717fd6a007aaac629f804cf3f42447bf4dd8d10323",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00007-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 1896341693,
+      "sha256": "b0c62fdcb51da61ba663200230641d47d1f9fa08444336ac45124be0b16696ba",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00008-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 2529808650,
+      "sha256": "203693c7c0b8642c4181fc7a9e90630c8d1db308eb4ad2f91c15c7c321609f2c",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00009-of-00011.parquet",
+      "rows": 1125000,
+      "bytes": 2775383707,
+      "sha256": "1f31fcbfff51c311d81bb110dc0f253b70fd2992f08300101452415f5ef7e267",
+      "split": "train"
+    },
+    {
+      "filename": "data/train/train-00010-of-00011.parquet",
+      "rows": 19408,
+      "bytes": 42750853,
+      "sha256": "2011b96b14e417c351a6e86d5ff2b22802d2ef697fb2cf073b9f0550d624f6a8",
+      "split": "train"
+    },
+    {
+      "filename": "data/validation/validation-00000-of-00001.parquet",
+      "rows": 354037,
+      "bytes": 570669729,
+      "sha256": "4d9b37d703f51c0cc31ad788b92b656417181f399946ee4f951e9bc6df204ef4",
+      "split": "validation"
+    }
+  ],
+  "shard_total_bytes": 22019379355,
+  "exact_row_identity_and_order_check": {
+    "train": {
+      "source_ordered_ids_sha256": "a8e63966d9616452d6e5c38a8bfd82c85f56fba03bdb500d825919ca4fbca7b3",
+      "shards_ordered_ids_sha256": "a8e63966d9616452d6e5c38a8bfd82c85f56fba03bdb500d825919ca4fbca7b3",
+      "rows": 11269408,
+      "match": true
+    },
+    "validation": {
+      "source_ordered_ids_sha256": "e551cd840c4942520938ebb88f175294051c5545b68a040784342613e94070c3",
+      "shards_ordered_ids_sha256": "e551cd840c4942520938ebb88f175294051c5545b68a040784342613e94070c3",
+      "rows": 354037,
+      "match": true
+    },
+    "test": {
+      "source_ordered_ids_sha256": "03fc50509409232351ff7566e6e1205600a01933f0d4066a68da1d726b3a3995",
+      "shards_ordered_ids_sha256": "03fc50509409232351ff7566e6e1205600a01933f0d4066a68da1d726b3a3995",
+      "rows": 354635,
+      "match": true
+    }
+  },
+  "schema": [
+    {
+      "name": "id",
+      "type": "string",
+      "nullable": false
+    },
+    {
+      "name": "source",
+      "type": "string",
+      "nullable": false
+    },
+    {
+      "name": "decision_type",
+      "type": "string",
+      "nullable": false
+    },
+    {
+      "name": "status",
+      "type": "string",
+      "nullable": false
+    },
+    {
+      "name": "unsupported_reason",
+      "type": "string",
+      "nullable": true
+    },
+    {
+      "name": "content_hash",
+      "type": "string",
+      "nullable": false
+    },
+    {
+      "name": "state",
+      "type": "struct<system: string, user_goal: string, history: list<element: struct<role: string, payload_json: string not null>> not null, environment_json: string>",
+      "nullable": false
+    },
+    {
+      "name": "candidates",
+      "type": "list<element: struct<id: string not null, name: string not null, description: string, parameters_json: string not null, metadata_json: string not null>>",
+      "nullable": false
+    },
+    {
+      "name": "target",
+      "type": "struct<candidate_id: string, action_name: string not null, arguments_json: string, label_json: string>",
+      "nullable": true
+    },
+    {
+      "name": "ordered_targets",
+      "type": "list<element: struct<candidate_id: string, action_name: string not null, arguments_json: string, label_json: string>>",
+      "nullable": false
+    },
+    {
+      "name": "labels",
+      "type": "struct<complete: bool, trajectory_success: bool, value_target: double, reward: double, score: double, source_pass_rate: double, outcome_evidence: string>",
+      "nullable": false
+    },
+    {
+      "name": "provenance",
+      "type": "struct<dataset_id: string not null, dataset_revision: string not null, source_config: string not null, raw_split: string not null, source_row_index: int64 not null, source_identity: string not null, trajectory_id: string not null, step_index: int64 not null, decision_ordinal: int64 not null, adapter_version: string not null, license_metadata_json: string not null, source_metadata_json: string not null>",
+      "nullable": false
+    },
+    {
+      "name": "training",
+      "type": "struct<choice_eligible: bool not null, completion_eligible: bool not null, value_eligible: bool not null, score_eligible: bool not null, boolean_eligible: bool not null, arguments_eligible: bool not null, use_for_bc: bool not null, use_for_value: bool not null, supervision_evidence: string not null, choice_ineligible_reason: string, quality_weight: double not null>",
+      "nullable": false
+    },
+    {
+      "name": "training_split",
+      "type": "string",
+      "nullable": true
+    }
+  ],
+  "method": "Rows were routed in source order by existing training_split and written as unmodified Arrow table slices. Every shard was fully re-read with PyArrow during build to validate file SHA256, schema, row count, and split purity. This finalization independently re-reads ID/split columns and verifies exact ordered unique-record identity against the canonical source."
+}

source_revisions.json

@@ -0,0 +1,32 @@
+{
+  "dataset": "samatv256/jev-decisions-v1",
+  "canonical_snapshot_cutoff": "2026-09-22T14:43:30.367628Z",
+  "canonical_source_revision_pins": {
+    "nvidia/Nemotron-SFT-Agentic-v2": "7c804833427f633ccd53b582dbf02525fd680f78",
+    "nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1": "9643c8103d7bfbc2d7fc4d15991d6739c612ff58",
+    "nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1": "cd26a12f54348e02a832d85f9da86621cd5dca4a",
+    "nvidia/Open-SWE-Traces": "233a7053326695357cce95405cc3df42239b0609"
+  },
+  "card_license_metadata_rechecked_at_publication": {
+    "checked_date": "2026-09-22",
+    "nvidia/Nemotron-SFT-Agentic-v2": {
+      "card_revision": "7c804833427f633ccd53b582dbf02525fd680f78",
+      "licenses": ["cc-by-4.0", "apache-2.0", "mit"]
+    },
+    "nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1": {
+      "card_revision": "9643c8103d7bfbc2d7fc4d15991d6739c612ff58",
+      "licenses": ["cc-by-4.0"]
+    },
+    "nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1": {
+      "card_revision": "cd26a12f54348e02a832d85f9da86621cd5dca4a",
+      "licenses": ["cc-by-4.0", "apache-2.0", "mit"]
+    },
+    "nvidia/Open-SWE-Traces": {
+      "card_revision": "f8fb5b3d2c787f85f8a00f5fe04fe3f1a11088ef",
+      "licenses": ["cc-by-4.0"],
+      "additional_card_licenses": ["mit", "apache-2.0", "bsd-2-clause", "bsd-3-clause"],
+      "row_level_license_field": "repository SPDX license"
+    }
+  },
+  "revision_note": "canonical_source_revision_pins record the content revisions used when building the corpus. Card revisions record the repositories observed during the pre-publication license check. Open-SWE v1.0 data revision differs from current main-card revision."
+}