| { |
| "checkpoint": { |
| "adaptation": "full", |
| "global_step": 600, |
| "records_seen": 96000, |
| "sha256": "020f76c434acb7652adcf0bc09f26c4605a00e7fe4462b6f71cea7f91438ebd5", |
| "weights": "online" |
| }, |
| "dataset_fingerprint": "d19f485febc970c3bd61b0ed4259c7229658c3e724c35817ee3806f66e614663", |
| "metrics": { |
| "completed": 444, |
| "completion_rate": 0.8671875, |
| "correct": 237, |
| "episodes": 512, |
| "final_answer_accuracy": 0.462890625, |
| "mean_calls": 6.55078125, |
| "oracle_action_exact_rate": 0.8631660631143301, |
| "oracle_plan_exact": 230, |
| "oracle_plan_exact_rate": 0.44921875, |
| "syntax_valid_turn_rate": 0.9824107604759441 |
| }, |
| "previously_unobserved_suffix": { |
| "correct": 208, |
| "episodes": 448, |
| "final_answer_accuracy": 0.4642857142857143, |
| "meaning": "Episodes 64..511 were not used for staged observation before the final sealed run.", |
| "start_index": 64 |
| }, |
| "report_scope": "aggregate_only; sealed prompts and trajectories withheld", |
| "sampling": { |
| "base_seed": 710001, |
| "sequence_length": 256, |
| "steps": 64 |
| }, |
| "schema_version": "sedd_public_aggregate_evaluation_v1", |
| "selection_range": { |
| "start": 0, |
| "stop": 512 |
| }, |
| "source_report": { |
| "included_in_public_repository": false, |
| "sha256": "7e434e1ecc607540a94d39239ae689dbc88374c23dcfc16f579225d30e6d54e6", |
| "size_bytes": 2953834 |
| }, |
| "split": "extrapolate" |
| } |
|
|
