| --- |
| pretty_name: SkillJit Case Study - Input Collections |
| language: |
| - en |
| - zh |
| license: other |
| license_name: mixed-source-terms |
| license_link: LICENSE |
| task_categories: |
| - other |
| tags: |
| - benchmark |
| - agent |
| - skill-learning |
| - private-staging |
| configs: |
| - config_name: 01-financial-analysis |
| data_files: |
| - split: instances |
| path: data/01-financial-analysis.parquet |
| - config_name: 03-court-form-filling |
| data_files: |
| - split: instances |
| path: data/03-court-form-filling.parquet |
| - config_name: 04-exam-block-sequencing |
| data_files: |
| - split: instances |
| path: data/04-exam-block-sequencing.parquet |
| - config_name: 05-hotel-book-room |
| data_files: |
| - split: instances |
| path: data/05-hotel-book-room.parquet |
| - config_name: 06-healthcare-submit-claim |
| data_files: |
| - split: instances |
| path: data/06-healthcare-submit-claim.parquet |
| - config_name: 07-content-flagging |
| data_files: |
| - split: instances |
| path: data/07-content-flagging.parquet |
| - config_name: 08-customer-service |
| data_files: |
| - split: instances |
| path: data/08-customer-service.parquet |
| - config_name: 09-pr-functional-risk-review |
| data_files: |
| - split: instances |
| path: data/09-pr-functional-risk-review.parquet |
| - config_name: 10-citation-check |
| data_files: |
| - split: instances |
| path: data/10-citation-check.parquet |
| --- |
| |
| # SkillJit case-study input collections |
|
|
| **Local, private-upload staging package; not a cleared public release.** |
| There are 900 expanded inputs: nine task configurations, 100 inputs each. |
| DBSCAN, separate historical pilots and original benchmark example collections |
| are excluded. Finance's ten retained pilot queries are already part of its |
| 100-input collection and are not counted again. |
|
|
| | Configuration and bilingual task document | Inputs | Maximum LLM calls | Timeout (seconds) | |
| |---|---:|---:|---:| |
| | [01-financial-analysis](docs/tasks/01-financial-analysis.md) | 100 | 40 | 900 | |
| | [03-court-form-filling](docs/tasks/03-court-form-filling.md) | 100 | 40 | 900 | |
| | [04-exam-block-sequencing](docs/tasks/04-exam-block-sequencing.md) | 100 | 60 | 1800 | |
| | [05-hotel-book-room](docs/tasks/05-hotel-book-room.md) | 100 | 20 | 300 | |
| | [06-healthcare-submit-claim](docs/tasks/06-healthcare-submit-claim.md) | 100 | 25 | 300 | |
| | [07-content-flagging](docs/tasks/07-content-flagging.md) | 100 | 20 | 300 | |
| | [08-customer-service](docs/tasks/08-customer-service.md) | 100 | 35 | 600 | |
| | [09-pr-functional-risk-review](docs/tasks/09-pr-functional-risk-review.md) | 100 | 60 | 1800 | |
| | [10-citation-check](docs/tasks/10-citation-check.md) | 100 | 40 / 80 / 150 | 900 / 1800 / 3600 | |
|
|
| ## Episode execution budgets |
|
|
| [benchmark_config.json](benchmark_config.json) is the versioned source of |
| call/time limits. Each row exports its resolved `execution_budget` with |
| `max_llm_calls` and `timeout_sec`. Citation values above are ordered by |
| 1-50, 51-200 and 201+ bibliography entries; select the tier before execution. |
| DBSCAN also has a configured budget but remains excluded from this package. |
|
|
| One call is one model-request attempt, including retries, subagents and |
| simulated users; parallel requests share one counter. Completion on the last |
| allowed call is permitted, but dispatching another is not. The episode timeout |
| includes LLM/network waiting, tools and CPU computation, excluding setup/reset, |
| grading and between-episode skill updates. An unfinished episode that requires |
| another call or reaches its deadline fails as `llm_call_limit_exceeded` or |
| `timeout`; stop outstanding work and ignore late results. Tokens are measured, |
| not capped. See the [full execution policy](docs/final_list.md#episode-execution-budgets). |
| These are our own starting limits, not upstream defaults or measured runtimes. |
| This package exports the policy; a consuming agent runner must enforce it. |
|
|
| ## Load with standard Hugging Face Datasets |
|
|
| ```python |
| from datasets import load_dataset |
| ds = load_dataset("YOUR_NAMESPACE/YOUR_DATASET", |
| name="10-citation-check", split="instances") |
| ``` |
|
|
| `instances` is the only HF split/container for every configuration. It contains |
| all 100 inputs; there is no training/validation/test allocation. |
| For local use, replace the repository ID with this package's directory. |
|
|
| ## Rows and file bundles |
|
|
| Each Parquet row contains an episode ID, instruction, an exact input-file |
| inventory, archive paths/prefixes, task-document path, reviewer metadata as JSON, |
| reference status, an optional bibliography entry count and the execution budget. |
|
|
| - `agent.tar.gz`: only episode input files, under their original repository paths. |
| - `support.tar.gz`: shared mounts, initial guidance and available support/notice files. |
| - `reviewer.tar.gz`: backend states, references, acquisition evidence and review-only files. |
| - `inventories/*.json`: exact archive member paths, SHA-256 hashes, sizes and executable bits. |
| - `provenance/`: original collection manifests, indexes, results and specifications. |
| - `docs/source/`: verbatim task-document snapshots, not a relocated runnable workspace. |
|
|
| Archives retain original repository-relative member paths. The unmodified |
| `data_mount.json` files resolve when the relevant bundles are unpacked under |
| the same workspace root. Finance's fixed SEC archives and the court blank |
| PDF are included as necessary shared mounts, not as additional episodes. |
|
|
| Only mount one selected episode's input files and its legitimate shared |
| support into an agent environment. Never mount reviewer bundles, Parquet |
| metadata, other episodes, private references or generation tools as agent |
| input. The word `private` in source paths denotes a visibility boundary, |
| not an access-control feature on a public HF repository. |
|
|
| ## Offline output evaluation |
|
|
| The [offline evaluator](benchmark_evaluation/README.md) is bundled as |
| `benchmark_evaluation/`. It grades outputs against expanded-case references, |
| not the original benchmarks' fixed answer keys: |
|
|
| ```bash |
| python -m pip install -r benchmark_evaluation/requirements.txt |
| python -B -m benchmark_evaluation --root /path/to/unpacked-workspace \ |
| --task 01-financial-analysis --episode 001 --submission /path/to/outputs \ |
| --result /path/to/grading/result.json --reward /path/to/grading/reward.txt |
| ``` |
|
|
| Unpack the task's agent/support/reviewer bundles into the grader workspace, |
| preserving member paths, and provide its `benchmark_config.json`. Never expose |
| reviewer material or grading results to the agent. Hotel/claim grading requires |
| a trusted runner-owned final-state/call/time record. Other output checks can |
| run without it, but report that budget compliance was not checked. |
| PR reviews require independent adjudication and remain `ungradable` without it. |
| Citation checks accept input-bound audited or explicitly authorized source-trust |
| labels; missing labels return `ungradable` with a null reward. The evaluator |
| documents task-specific scope, prose/semantic limitations, and status/exit codes. |
| The package `verify` command below checks integrity, not task correctness. |
|
|
| This is a file-backed dataset with offline output checks, not a certified agent |
| runtime or a universal semantic correctness oracle. |
| Initial guidance and available benchmark helpers are preserved. Skill-local |
| `scripts/`, `references/`, `assets/` and notices are included under `skills/` |
| and in support archives at their original paths. Native support Python modules |
| and tool schemas are also included. Original files are unchanged: run helpers |
| from their skill directory, and unpack support archives for workspace-relative |
| guidance links. Third-party packages, benchmark runtimes (such as SOPBench's |
| `env` package) and services still need installation/configuration. SOP tool |
| modules require a controlled backend; private lookup data is not agent support. |
| Citation guidance's optional `scripts/generate_schematic.py` belongs to a |
| separate skill absent from the checked-out benchmark and is not bundled. |
| Downloads must not execute bundled PR code or TeX. |
|
|
| ## Labels and source limitations |
|
|
| **No fake references were added to the 100 citation inputs.** Their author |
| bibliographies are byte-identical, with 7,356 entries total and natural sizes |
| of 9-972 entries. Where `reference_status` is `assumed_all_real`, private labels |
| use `fake_citations: []` under the user's explicit source-trust decision, not |
| independent auditing or certified gold. An always-empty answer passes those |
| cases; the all-negative collection cannot measure fake-reference detection. |
| Cases without labels retain `reference_status: unknown`. |
| The source-provider revision and file hashes are pinned; original arXiv |
| source versions and per-file/version redistribution rights are not certified. |
|
|
| Recorded references in other tasks are construction/replay evidence, not |
| uniformly independently certified gold labels or agent success rates. |
| Task-specific limits, licensing notes and generation methods are in the |
| bilingual task documents and source snapshots. |
|
|
| ## Integrity and upload |
|
|
| Install `requirements.txt`, then run: |
|
|
| ```bash |
| python tools/package_hf_dataset.py verify --output . --load |
| python tools/package_hf_dataset.py upload --output . --repo-id YOUR_NAMESPACE/YOUR_DATASET |
| ``` |
|
|
| The upload command defaults to a network-free dry run. After reviewing rights, |
| privacy, bundled source terms and reviewer-data exposure, explicitly add: |
|
|
| ```bash |
| --execute --rights-reviewed --reviewer-material-reviewed |
| ``` |
|
|
| Execution creates a private HF dataset or requires an existing private dataset. |
| An existing public repository is refused. A private target is not a substitute |
| for permission: do not execute until every included asset may be uploaded. |
| The exporter did not read HF credentials, upload data or certify release rights. |
|
|
