Download experiment-v2-precommit.json from Aditya0427/ToolBoundary-authorization-v2: direct link, hf CLI and curl.
- Browser
- Download file 1.03 kB
-
https://huggingface.co/datasets/Aditya0427/ToolBoundary-authorization-v2/resolve/main/experiment-v2-precommit.json
- Command line
-
hf download hf://datasets/Aditya0427/ToolBoundary-authorization-v2/experiment-v2-precommit.json
-
curl -L -o experiment-v2-precommit.json https://huggingface.co/datasets/Aditya0427/ToolBoundary-authorization-v2/resolve/main/experiment-v2-precommit.json
1.03 kB
| { | |
| "phase": "v2_completion_only", | |
| "steps": 60, | |
| "training_cases": 240, | |
| "evaluation_cases": 48, | |
| "created_before_any_v2_model_outputs": true, | |
| "changes_from_failed_pilot": [ | |
| "explicit raw JSON/no-Markdown instruction applied equally to baseline and adapter", | |
| "completion-only loss", | |
| "new separate training/evaluation IDs and heldout wording" | |
| ], | |
| "hashes": { | |
| "training-v2.json": "31055795ff57b7bc7c034afb69b65f1126e8a6933bad20e5bc871bb11b456c62", | |
| "evaluation-v2.json": "534580cf3acc1653c9bd4f0ba0611a0e807fd2df64d084e895caa93c19f0c9fa", | |
| "toolboundary.py": "3706af4fa163554e2c5ee84170a06aff028501012ee4aee43ce4e9b874159994", | |
| "run_experiment.py": "e7cc0e757d7d8d60f15cf1ff01dd91069cd2092435981c22420d690c33c15406" | |
| }, | |
| "metrics": [ | |
| "strict JSON validity", | |
| "exact task action", | |
| "unsafe write attempts", | |
| "unnecessary approvals" | |
| ], | |
| "claim_boundary": "Inspectable synthetic corpus; no independent real user validation or blind external evaluation" | |
| } |