Datasets:
Zero Proof AI
character training on the OpenAI Model Spec: train, holdout, eval (zeroproof 0.24)
c044c8c verified | { | |
| "source": "zeroproof 0.24, examples/character, live run 2026-09-14", | |
| "splits": { | |
| "train": { | |
| "rows": 60, | |
| "prompts": 15, | |
| "pass_rate": 0.7167, | |
| "platform_dataset": "ds_004747715e398d3b", | |
| "kinds": { | |
| "example": 60 | |
| } | |
| }, | |
| "holdout": { | |
| "rows": 144, | |
| "prompts": 36, | |
| "pass_rate": 0.8819, | |
| "platform_dataset": "ds_5bcd0ddf95f4896c", | |
| "kinds": { | |
| "adversarial": 120, | |
| "control": 24 | |
| } | |
| }, | |
| "eval": { | |
| "rows": 35, | |
| "prompts": 15, | |
| "pass_rate": 0.6857, | |
| "platform_dataset": "ds_a6e6f830d7e61b48", | |
| "kinds": { | |
| "spec": 35 | |
| } | |
| } | |
| } | |
| } |