typical-small-preview / eval_wf.json
guychuk's picture
Publish release artefacts + model card + MANIFEST
1c85040 verified
Raw History Blame Contribute Delete
5.93 kB
{
"run": "runs/nc_v3_tap20_wf",
"mode": "native",
"max_state": 4096,
"limit": 500,
"eval": {
"wf_heldout_choice": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.019871275152449026,
"brier": 0.010466870652060482,
"jsd": 0.005765777761771096,
"acc": 0.994,
"acc_k": 0.9937238493723849,
"ece": 0.0063112200498580615,
"sel_acc80": 1.0,
"conf_wrong": 0.0,
"coverage": 0.956,
"false_abstain": 0.0,
"null_recall": 1.0,
"auroc_null": 1.0,
"auroc_conf": 1.0
}
},
"wf_heldout_noul": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.853929875950949,
"brier": 0.4878031580677193,
"jsd": 0.17909886291906896,
"acc": 0.682,
"acc_k": 0.682,
"ece": 0.1986952285766601,
"sel_acc80": 0.715,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"wf_heldout_score": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 2.068675406193429,
"brier": 0.7677929317097167,
"jsd": 0.2869982566550366,
"acc": 0.514,
"acc_k": 0.514,
"ece": 0.35427253764867783,
"sel_acc80": 0.5725,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"wf_heldout_style": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.2222467604358654,
"brier": 0.1323241858578231,
"jsd": 0.06029804682074418,
"acc": 0.904,
"acc_k": 0.901840490797546,
"ece": 0.019670174241066027,
"sel_acc80": 0.975,
"conf_wrong": 0.0,
"coverage": 0.978,
"false_abstain": 0.0,
"null_recall": 1.0,
"auroc_null": 1.0,
"auroc_conf": 1.0
}
},
"wf_rubric_flip": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.7009927043024266,
"brier": 0.40828858133650053,
"jsd": 0.15320314054370474,
"acc": 0.736,
"acc_k": 0.736,
"ece": 0.1480658851861954,
"sel_acc80": 0.7825,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
},
"flip": {
"n_pairs": 250,
"flip_rate": 0.504,
"both_correct": 0.488,
"acc": 0.736
}
},
"wf_rubric_shuffled": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 4.60017100450565,
"brier": 1.0415848067241245,
"jsd": 0.3751870274955275,
"acc": 0.386,
"acc_k": 0.386,
"ece": 0.5058854785561562,
"sel_acc80": 0.395,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"cua_s1_forms_demo": {
"n": 196,
"suffix_overflow": 0,
"raw": {
"nll": 0.00011844338656585847,
"brier": 3.8039107351542767e-06,
"jsd": 4.095999434756056e-05,
"acc": 1.0,
"acc_k": 1.0,
"ece": 0.00011747923432559038,
"sel_acc80": 1.0,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"cua_s1_forms_test": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.012289381821878892,
"brier": 0.004094414644142791,
"jsd": 0.001583947825507208,
"acc": 0.998,
"acc_k": 0.998,
"ece": 0.002032796382904006,
"sel_acc80": 1.0,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"jev4b_adversarial": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.3490933191998588,
"brier": 0.09980970638364663,
"jsd": 0.040403990439927887,
"acc": 0.94,
"acc_k": 0.94,
"ece": 0.04506945902109154,
"sel_acc80": 0.965,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"jevlogs_triage": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.3044197695283073,
"brier": 0.17586170974258888,
"jsd": 0.0972293910378737,
"acc": 0.832,
"acc_k": 0.832,
"ece": 0.23506193399429315,
"sel_acc80": 1.0,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"mind2web_choice": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 1.7049515437075962,
"brier": 0.7232633764663069,
"jsd": 0.3240744768308897,
"acc": 0.438,
"acc_k": 0.438,
"ece": 0.21502542391419413,
"sel_acc80": 0.5025,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"pagerduty_trigger": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.6816042798599857,
"brier": 0.3284166395830621,
"jsd": 0.14193714639027197,
"acc": 0.762,
"acc_k": 0.762,
"ece": 0.10664146476984024,
"sel_acc80": 0.865,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"systemone_lite_hard": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 0.2668189883849799,
"brier": 0.15873402095268352,
"jsd": 0.06530097270690566,
"acc": 0.888,
"acc_k": 0.888,
"ece": 0.03896239256858824,
"sel_acc80": 0.9625,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"tree_choice_cap": {
"n": 500,
"suffix_overflow": 125,
"raw": {
"nll": 2.243535988875024,
"brier": 0.6146624018620521,
"jsd": 0.34488693403770676,
"acc": 0.494,
"acc_k": 0.538,
"ece": 0.07969999838478839,
"sel_acc80": 0.6175,
"coverage": 0.872,
"false_abstain": 0.128,
"null_recall": "n/a"
}
},
"typed_decisions_test": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 2.4008272248719806,
"brier": 0.5420563405144715,
"jsd": 0.2209605208544786,
"acc": 0.276,
"acc_k": 0.276,
"ece": 0.5269120357036591,
"sel_acc80": 0.2925,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
},
"typed_decisions_train": {
"n": 500,
"suffix_overflow": 0,
"raw": {
"nll": 2.5297219974097365,
"brier": 0.5571985794974775,
"jsd": 0.2336201678966866,
"acc": 0.288,
"acc_k": 0.288,
"ece": 0.511909948706627,
"sel_acc80": 0.2875,
"coverage": 1.0,
"false_abstain": 0.0,
"null_recall": "n/a"
}
}
}
}