membrane-temp / docs /eval /base_vs_trained /base_vs_trained_summary.json
luffy-orf
submission: Membrane docs, training/eval artifacts, HF jobs scripts; docs plots as SVG only (HF Space git binary hook)
9c49c6e
Raw
History Blame Contribute Delete
1.39 kB
{
"checkpoint_run": "continue_deep_seed_5821_steps_1500",
"samples_per_task": 4,
"tasks": [
"dyad_must_refuse_v1",
"dyad_must_comply_v1",
"dyad_must_refuse_long_v1",
"triad_must_refuse_v1"
],
"base": {
"mean_reward": {
"dyad_must_refuse_v1": 0.0,
"dyad_must_comply_v1": 0.0,
"dyad_must_refuse_long_v1": 0.0,
"triad_must_refuse_v1": 0.0
},
"valid_jsonl_rate": {
"dyad_must_refuse_v1": 0.0,
"dyad_must_comply_v1": 0.0,
"dyad_must_refuse_long_v1": 0.0,
"triad_must_refuse_v1": 0.0
},
"commit_rate": {
"dyad_must_refuse_v1": 0.0,
"dyad_must_comply_v1": 0.0,
"dyad_must_refuse_long_v1": 0.0,
"triad_must_refuse_v1": 0.0
},
"overall_mean_reward": 0.0
},
"trained": {
"mean_reward": {
"dyad_must_refuse_v1": 0.988,
"dyad_must_comply_v1": 0.435,
"dyad_must_refuse_long_v1": 0.9957142857142858,
"triad_must_refuse_v1": 0.6416666666666667
},
"valid_jsonl_rate": {
"dyad_must_refuse_v1": 1.0,
"dyad_must_comply_v1": 1.0,
"dyad_must_refuse_long_v1": 1.0,
"triad_must_refuse_v1": 1.0
},
"commit_rate": {
"dyad_must_refuse_v1": 1.0,
"dyad_must_comply_v1": 1.0,
"dyad_must_refuse_long_v1": 1.0,
"triad_must_refuse_v1": 1.0
},
"overall_mean_reward": 0.7650952380952382
}
}