kev-0.8b-code-verify-v2 / run_eval.json
jtatman's picture
Upload code-verify-08b-v2
2d24402 verified
Raw History Blame Contribute Delete
3.87 kB
{
"development": {
"baseline kev": {
"auroc": 0.8941919191919192,
"acc": 0.9553571428571429,
"brier": 0.04663790840267551,
"ece": 0.054507326705475256,
"cov@2%": 0.12946428571428573,
"cov@5%": 0.84375,
"cov@10%": 0.8928571428571429
},
"fine-tuned kev": {
"auroc": 0.9127209595959596,
"acc": 0.9553571428571429,
"brier": 0.04322990868327309,
"ece": 0.026368809832356274,
"cov@2%": 0.22321428571428573,
"cov@5%": 0.8258928571428571,
"cov@10%": 0.8928571428571429
},
"evidence score": {
"auroc": 0.9590277777777778,
"acc": 0.9553571428571429,
"brier": 0.04669189453125,
"ece": 0.07882254464285716,
"cov@2%": 0.6852678571428571,
"cov@5%": 0.8058035714285714,
"cov@10%": 0.8794642857142857
},
"mean(evidence, baseline)": {
"auroc": 0.9642992424242425,
"acc": 0.9553571428571429,
"brier": 0.04325954805619029,
"ece": 0.06962798473367474,
"cov@2%": 0.7433035714285714,
"cov@5%": 0.8125,
"cov@10%": 0.8928571428571429
},
"mean(evidence, fine-tuned)": {
"auroc": 0.9620580808080809,
"acc": 0.953125,
"brier": 0.041460332654909955,
"ece": 0.05010392415948695,
"cov@2%": 0.7232142857142857,
"cov@5%": 0.8125,
"cov@10%": 0.8928571428571429
}
},
"test_unseen_coder": {
"baseline kev": {
"auroc": 0.8957307060755336,
"acc": 0.9154929577464789,
"brier": 0.08782003066740127,
"ece": 0.10255376683858186,
"cov@2%": 0.09859154929577464,
"cov@5%": 0.09859154929577464,
"cov@10%": 0.30985915492957744
},
"fine-tuned kev": {
"auroc": 0.9490968801313628,
"acc": 0.9154929577464789,
"brier": 0.07603090115173004,
"ece": 0.06278563419018489,
"cov@2%": 0.36619718309859156,
"cov@5%": 0.4084507042253521,
"cov@10%": 0.5774647887323944
},
"evidence score": {
"auroc": 0.9433497536945813,
"acc": 0.9154929577464789,
"brier": 0.08483714788732394,
"ece": 0.1338028169014084,
"cov@2%": 0.15492957746478872,
"cov@5%": 0.15492957746478872,
"cov@10%": 0.5915492957746479
},
"mean(evidence, baseline)": {
"auroc": 0.9417077175697866,
"acc": 0.9154929577464789,
"brier": 0.08107140750327455,
"ece": 0.11410879479547414,
"cov@2%": 0.2535211267605634,
"cov@5%": 0.2535211267605634,
"cov@10%": 0.6056338028169014
},
"mean(evidence, fine-tuned)": {
"auroc": 0.9482758620689655,
"acc": 0.9154929577464789,
"brier": 0.07402588517908092,
"ece": 0.1054580941083675,
"cov@2%": 0.30985915492957744,
"cov@5%": 0.323943661971831,
"cov@10%": 0.6056338028169014
}
},
"test_unseen_coder_all_tasks": {
"baseline kev": {
"auroc": 0.9353021978021978,
"acc": 0.8995901639344263,
"brier": 0.09627491548962074,
"ece": 0.11438050809440908,
"cov@2%": 0.22540983606557377,
"cov@5%": 0.2766393442622951,
"cov@10%": 0.47950819672131145
},
"fine-tuned kev": {
"auroc": 0.9695741758241758,
"acc": 0.9364754098360656,
"brier": 0.05493458662718413,
"ece": 0.04307024866475148,
"cov@2%": 0.2192622950819672,
"cov@5%": 0.5061475409836066,
"cov@10%": 0.6372950819672131
},
"evidence score": {
"auroc": 0.943543956043956,
"acc": 0.9057377049180327,
"brier": 0.09111320255251247,
"ece": 0.09400301997638061,
"cov@2%": 0.0,
"cov@5%": 0.20901639344262296,
"cov@10%": 0.6188524590163934
},
"mean(evidence, baseline)": {
"auroc": 0.9518028846153846,
"acc": 0.9057377049180327,
"brier": 0.08850434302950522,
"ece": 0.09898749653604542,
"cov@2%": 0.11475409836065574,
"cov@5%": 0.375,
"cov@10%": 0.6086065573770492
},
"mean(evidence, fine-tuned)": {
"auroc": 0.9614354395604395,
"acc": 0.9344262295081968,
"brier": 0.06332625202060693,
"ece": 0.07127569574009193,
"cov@2%": 0.18032786885245902,
"cov@5%": 0.38729508196721313,
"cov@10%": 0.6372950819672131
}
}
}