Ines-1 / eval /fullcase_report.json
Endikavi's picture
Ines-1 RC1 (private staging; release commit b7f5644)
61b6fb9 verified
Raw History Blame Contribute Delete
16.7 kB
{
"v2|typed_en|single-question": {
"n": 2000,
"correct": 1510,
"accuracy": 0.755,
"nll": 0.6463324192750852,
"kl": 0.13270933329166798,
"ce_soft": 0.9000678889159969,
"brier": 0.07002914815029006,
"brier_mean_k": 0.02004525517631372,
"brier_onehot": 0.3676075399939686,
"ece10": 0.12290431407094005,
"mean_conf": 0.6332694019079208,
"by_type": {
"choice": {
"correct": 411,
"n": 600,
"accuracy": 0.685
},
"noul": {
"correct": 503,
"n": 600,
"accuracy": 0.8383333333333334
},
"score": {
"correct": 596,
"n": 800,
"accuracy": 0.745
}
},
"ci95_cases": [
0.7335,
0.7765
],
"score_mae": {
"mae_vs_gold_expectation": 0.24199143103868068,
"mae_vs_gold_label": 0.4274223285297194,
"n": 800
}
},
"v2|typed_en|full-case context": {
"n": 2000,
"correct": 857,
"accuracy": 0.4285,
"nll": 1.2097812919002324,
"kl": 0.4959933924254989,
"ce_soft": 1.2633519480498279,
"brier": 0.26928889266663847,
"brier_mean_k": 0.08697066222849263,
"brier_onehot": 0.6970586865702736,
"ece10": 0.11118127906322481,
"mean_conf": 0.5374737640917301,
"by_type": {
"choice": {
"correct": 198,
"n": 600,
"accuracy": 0.33
},
"noul": {
"correct": 298,
"n": 600,
"accuracy": 0.49666666666666665
},
"score": {
"correct": 361,
"n": 800,
"accuracy": 0.45125
}
},
"ci95_cases": [
0.407,
0.45
],
"score_mae": {
"mae_vs_gold_expectation": 0.5068433761781231,
"mae_vs_gold_label": 0.6975457973429002,
"n": 800
}
},
"v2|typed_en|full-case joint": {
"n": 2000,
"correct": 706,
"accuracy": 0.353,
"nll": 1.217202288315301,
"kl": 0.4577465691470793,
"ce_soft": 1.2251051247714082,
"brier": 0.2549271664455238,
"brier_mean_k": 0.07846016753062504,
"brier_onehot": 0.7019905508599895,
"ece10": 0.12446078114211558,
"mean_conf": 0.4774607811421156,
"by_type": {
"choice": {
"correct": 193,
"n": 600,
"accuracy": 0.32166666666666666
},
"noul": {
"correct": 286,
"n": 600,
"accuracy": 0.4766666666666667
},
"score": {
"correct": 227,
"n": 800,
"accuracy": 0.28375
}
},
"ci95_cases": [
0.3315,
0.374
],
"score_mae": {
"mae_vs_gold_expectation": 0.7265793839785138,
"mae_vs_gold_label": 0.9440566271264106,
"n": 800
}
},
"v2|typed_en|single-question vs full-case context": {
"n_common": 2000,
"only_b_right": 144,
"only_a_right": 797,
"p_mcnemar_exact": 3.605359197992819e-110
},
"v2|typed_en|full-case context vs full-case joint": {
"n_common": 2000,
"only_b_right": 302,
"only_a_right": 453,
"p_mcnemar_exact": 4.334154965648315e-08
},
"v2|typed_es|single-question": {
"n": 1985,
"correct": 1482,
"accuracy": 0.7465994962216624,
"nll": 0.6550145421374155,
"kl": 0.13250811726703088,
"ce_soft": 0.9007772998807515,
"brier": 0.07028015084314279,
"brier_mean_k": 0.01999555773940268,
"brier_onehot": 0.3732510221786229,
"ece10": 0.1152207864471887,
"mean_conf": 0.6313787097744737,
"by_type": {
"choice": {
"correct": 401,
"n": 594,
"accuracy": 0.6750841750841751
},
"noul": {
"correct": 497,
"n": 597,
"accuracy": 0.8324958123953099
},
"score": {
"correct": 584,
"n": 794,
"accuracy": 0.7355163727959698
}
},
"ci95_cases": [
0.7249370277078085,
0.7677581863979849
],
"score_mae": {
"mae_vs_gold_expectation": 0.23905470773449672,
"mae_vs_gold_label": 0.4315963490568928,
"n": 794
}
},
"v2|typed_es|full-case context": {
"n": 1985,
"correct": 832,
"accuracy": 0.41914357682619646,
"nll": 1.2000239557178545,
"kl": 0.47456987091550584,
"ce_soft": 1.2428390535292264,
"brier": 0.26059998598083783,
"brier_mean_k": 0.08474363410961229,
"brier_onehot": 0.6933349301776697,
"ece10": 0.10654264947479258,
"mean_conf": 0.523837870787313,
"by_type": {
"choice": {
"correct": 236,
"n": 594,
"accuracy": 0.39730639730639733
},
"noul": {
"correct": 295,
"n": 597,
"accuracy": 0.49413735343383586
},
"score": {
"correct": 301,
"n": 794,
"accuracy": 0.3790931989924433
}
},
"ci95_cases": [
0.39848866498740554,
0.4397984886649874
],
"score_mae": {
"mae_vs_gold_expectation": 0.5519418643797106,
"mae_vs_gold_label": 0.7501145362577187,
"n": 794
}
},
"v2|typed_es|full-case joint": {
"n": 1985,
"correct": 673,
"accuracy": 0.33904282115869017,
"nll": 1.2040304828794082,
"kl": 0.44507417616530687,
"ce_soft": 1.2133433587790274,
"brier": 0.24717614322069642,
"brier_mean_k": 0.07750307326535547,
"brier_onehot": 0.6933235732686921,
"ece10": 0.11217208232176995,
"mean_conf": 0.4512149034804601,
"by_type": {
"choice": {
"correct": 167,
"n": 594,
"accuracy": 0.28114478114478114
},
"noul": {
"correct": 284,
"n": 597,
"accuracy": 0.47571189279731996
},
"score": {
"correct": 222,
"n": 794,
"accuracy": 0.2795969773299748
}
},
"ci95_cases": [
0.31738035264483627,
0.3602015113350126
],
"score_mae": {
"mae_vs_gold_expectation": 0.6897287760172528,
"mae_vs_gold_label": 0.921904983332085,
"n": 794
}
},
"v2|typed_es|single-question vs full-case context": {
"n_common": 1985,
"only_b_right": 119,
"only_a_right": 769,
"p_mcnemar_exact": 3.7714035666656705e-117
},
"v2|typed_es|full-case context vs full-case joint": {
"n_common": 1985,
"only_b_right": 254,
"only_a_right": 413,
"p_mcnemar_exact": 7.935736100989516e-10
},
"v2|telepatia_es|single-question": {
"n": 2000,
"correct": 1492,
"accuracy": 0.746,
"nll": 0.6768666282900727,
"kl": 0.137470178826373,
"ce_soft": 0.9048287344507019,
"brier": 0.07226937229140096,
"brier_mean_k": 0.020558110269673965,
"brier_onehot": 0.3838231707888498,
"ece10": 0.1303335434347391,
"mean_conf": 0.6159254869967699,
"by_type": {
"choice": {
"correct": 415,
"n": 600,
"accuracy": 0.6916666666666667
},
"noul": {
"correct": 502,
"n": 600,
"accuracy": 0.8366666666666667
},
"score": {
"correct": 575,
"n": 800,
"accuracy": 0.71875
}
},
"ci95_cases": [
0.7245,
0.768
],
"score_mae": {
"mae_vs_gold_expectation": 0.26117300274514005,
"mae_vs_gold_label": 0.44825553921662503,
"n": 800
}
},
"v2|telepatia_es|full-case context": {
"n": 2000,
"correct": 823,
"accuracy": 0.4115,
"nll": 1.189682075850548,
"kl": 0.4673731803374047,
"ce_soft": 1.2347317359617336,
"brier": 0.26037437617486775,
"brier_mean_k": 0.08499574423415869,
"brier_onehot": 0.6928821846712389,
"ece10": 0.1040707949474454,
"mean_conf": 0.5111226370260119,
"by_type": {
"choice": {
"correct": 199,
"n": 600,
"accuracy": 0.33166666666666667
},
"noul": {
"correct": 291,
"n": 600,
"accuracy": 0.485
},
"score": {
"correct": 333,
"n": 800,
"accuracy": 0.41625
}
},
"ci95_cases": [
0.39,
0.433
],
"score_mae": {
"mae_vs_gold_expectation": 0.5304797743702488,
"mae_vs_gold_label": 0.7196103962347843,
"n": 800
}
},
"v2|telepatia_es|full-case joint": {
"n": 2000,
"correct": 636,
"accuracy": 0.318,
"nll": 1.2328385029619342,
"kl": 0.4669040681698344,
"ce_soft": 1.2342626237941632,
"brier": 0.2603068636325369,
"brier_mean_k": 0.08182894089557148,
"brier_onehot": 0.7105703100725699,
"ece10": 0.15144768108427525,
"mean_conf": 0.46944768108427526,
"by_type": {
"choice": {
"correct": 125,
"n": 600,
"accuracy": 0.20833333333333334
},
"noul": {
"correct": 298,
"n": 600,
"accuracy": 0.49666666666666665
},
"score": {
"correct": 213,
"n": 800,
"accuracy": 0.26625
}
},
"ci95_cases": [
0.296,
0.34
],
"score_mae": {
"mae_vs_gold_expectation": 0.6871891529425792,
"mae_vs_gold_label": 0.913369414373301,
"n": 800
}
},
"v2|telepatia_es|single-question vs full-case context": {
"n_common": 2000,
"only_b_right": 126,
"only_a_right": 795,
"p_mcnemar_exact": 2.2596591083909646e-119
},
"v2|telepatia_es|full-case context vs full-case joint": {
"n_common": 2000,
"only_b_right": 261,
"only_a_right": 448,
"p_mcnemar_exact": 2.1310615747335922e-12
},
"v3|typed_en|single-question": {
"n": 2000,
"correct": 1528,
"accuracy": 0.764,
"nll": 0.6348000373513565,
"kl": 0.12133635641347341,
"ce_soft": 0.8886949120378023,
"brier": 0.0651098222513845,
"brier_mean_k": 0.019005062125487936,
"brier_onehot": 0.3611829084554942,
"ece10": 0.13159668615460393,
"mean_conf": 0.632403313845396,
"by_type": {
"choice": {
"correct": 429,
"n": 600,
"accuracy": 0.715
},
"noul": {
"correct": 503,
"n": 600,
"accuracy": 0.8383333333333334
},
"score": {
"correct": 596,
"n": 800,
"accuracy": 0.745
}
},
"ci95_cases": [
0.742,
0.7855
],
"score_mae": {
"mae_vs_gold_expectation": 0.23437388759128985,
"mae_vs_gold_label": 0.42270025751247886,
"n": 800
}
},
"v3|typed_en|full-case context": {
"n": 2000,
"correct": 831,
"accuracy": 0.4155,
"nll": 1.1946701619356137,
"kl": 0.48748056433424275,
"ce_soft": 1.2548391199585716,
"brier": 0.26584338076551384,
"brier_mean_k": 0.08537406918322844,
"brier_onehot": 0.6924497841014353,
"ece10": 0.11764439262449741,
"mean_conf": 0.5266776051074267,
"by_type": {
"choice": {
"correct": 153,
"n": 600,
"accuracy": 0.255
},
"noul": {
"correct": 295,
"n": 600,
"accuracy": 0.49166666666666664
},
"score": {
"correct": 383,
"n": 800,
"accuracy": 0.47875
}
},
"ci95_cases": [
0.3945,
0.4365
],
"score_mae": {
"mae_vs_gold_expectation": 0.5401024079043407,
"mae_vs_gold_label": 0.7221055129170417,
"n": 800
}
},
"v3|typed_en|full-case joint": {
"n": 2000,
"correct": 742,
"accuracy": 0.371,
"nll": 1.1821842679997898,
"kl": 0.43716691089620746,
"ce_soft": 1.2045254665205365,
"brier": 0.2458284892200497,
"brier_mean_k": 0.0777216018771812,
"brier_onehot": 0.6827152191861325,
"ece10": 0.08931016900390386,
"mean_conf": 0.44944229010492565,
"by_type": {
"choice": {
"correct": 172,
"n": 600,
"accuracy": 0.2866666666666667
},
"noul": {
"correct": 276,
"n": 600,
"accuracy": 0.46
},
"score": {
"correct": 294,
"n": 800,
"accuracy": 0.3675
}
},
"ci95_cases": [
0.349,
0.393
],
"score_mae": {
"mae_vs_gold_expectation": 0.6182189710583621,
"mae_vs_gold_label": 0.8307282497920095,
"n": 800
}
},
"v3|typed_en|single-question vs full-case context": {
"n_common": 2000,
"only_b_right": 150,
"only_a_right": 847,
"p_mcnemar_exact": 1.495086609422151e-118
},
"v3|typed_en|full-case context vs full-case joint": {
"n_common": 2000,
"only_b_right": 385,
"only_a_right": 474,
"p_mcnemar_exact": 0.0026569907313413405
},
"v3|typed_es|single-question": {
"n": 1985,
"correct": 1496,
"accuracy": 0.7536523929471033,
"nll": 0.6394842117466736,
"kl": 0.11898906052638147,
"ce_soft": 0.887258243140102,
"brier": 0.06475655764622247,
"brier_mean_k": 0.018860197074895686,
"brier_onehot": 0.36670251099112927,
"ece10": 0.12293764144107136,
"mean_conf": 0.6313106105219507,
"by_type": {
"choice": {
"correct": 415,
"n": 594,
"accuracy": 0.6986531986531986
},
"noul": {
"correct": 494,
"n": 597,
"accuracy": 0.8274706867671692
},
"score": {
"correct": 587,
"n": 794,
"accuracy": 0.739294710327456
}
},
"ci95_cases": [
0.7324937027707809,
0.7743073047858942
],
"score_mae": {
"mae_vs_gold_expectation": 0.23596503181925557,
"mae_vs_gold_label": 0.4296506547049947,
"n": 794
}
},
"v3|typed_es|full-case context": {
"n": 1985,
"correct": 859,
"accuracy": 0.4327455919395466,
"nll": 1.162280408384133,
"kl": 0.4496600929573715,
"ce_soft": 1.217929275571092,
"brier": 0.2505504103004089,
"brier_mean_k": 0.08182399037164234,
"brier_onehot": 0.6788902020159528,
"ece10": 0.09863890937203425,
"mean_conf": 0.5196653394690088,
"by_type": {
"choice": {
"correct": 211,
"n": 594,
"accuracy": 0.3552188552188552
},
"noul": {
"correct": 296,
"n": 597,
"accuracy": 0.49581239530988275
},
"score": {
"correct": 352,
"n": 794,
"accuracy": 0.4433249370277078
}
},
"ci95_cases": [
0.4105793450881612,
0.4544080604534005
],
"score_mae": {
"mae_vs_gold_expectation": 0.5462041908321706,
"mae_vs_gold_label": 0.7497829019000534,
"n": 794
}
},
"v3|typed_es|full-case joint": {
"n": 1985,
"correct": 763,
"accuracy": 0.38438287153652395,
"nll": 1.172466296474459,
"kl": 0.42073980961851604,
"ce_soft": 1.1890089922322367,
"brier": 0.23343222723881618,
"brier_mean_k": 0.07347218898601528,
"brier_onehot": 0.6736496081085933,
"ece10": 0.07710073747923152,
"mean_conf": 0.4243709911927768,
"by_type": {
"choice": {
"correct": 225,
"n": 594,
"accuracy": 0.3787878787878788
},
"noul": {
"correct": 286,
"n": 597,
"accuracy": 0.47906197654941374
},
"score": {
"correct": 252,
"n": 794,
"accuracy": 0.31738035264483627
}
},
"ci95_cases": [
0.36171284634760703,
0.4070528967254408
],
"score_mae": {
"mae_vs_gold_expectation": 0.6776316444522275,
"mae_vs_gold_label": 0.9095255444835055,
"n": 794
}
},
"v3|typed_es|single-question vs full-case context": {
"n_common": 1985,
"only_b_right": 137,
"only_a_right": 774,
"p_mcnemar_exact": 1.6584466757134775e-108
},
"v3|typed_es|full-case context vs full-case joint": {
"n_common": 1985,
"only_b_right": 320,
"only_a_right": 416,
"p_mcnemar_exact": 0.0004544874488207351
},
"v3|telepatia_es|single-question": {
"n": 2000,
"correct": 1526,
"accuracy": 0.763,
"nll": 0.6385563440375813,
"kl": 0.11925372995655213,
"ce_soft": 0.886612285580881,
"brier": 0.06393994905918735,
"brier_mean_k": 0.018569114373183912,
"brier_onehot": 0.36353173992639604,
"ece10": 0.13380355775356292,
"mean_conf": 0.6294715915620327,
"by_type": {
"choice": {
"correct": 438,
"n": 600,
"accuracy": 0.73
},
"noul": {
"correct": 498,
"n": 600,
"accuracy": 0.83
},
"score": {
"correct": 590,
"n": 800,
"accuracy": 0.7375
}
},
"ci95_cases": [
0.742,
0.785
],
"score_mae": {
"mae_vs_gold_expectation": 0.23628337060784574,
"mae_vs_gold_label": 0.42863585742670696,
"n": 800
}
},
"v3|telepatia_es|full-case context": {
"n": 2000,
"correct": 878,
"accuracy": 0.439,
"nll": 1.1470116522285891,
"kl": 0.4409125760498499,
"ce_soft": 1.2082711316741788,
"brier": 0.24666331201318742,
"brier_mean_k": 0.08078183830888044,
"brier_onehot": 0.671401023710438,
"ece10": 0.09246306705474854,
"mean_conf": 0.5134267296791076,
"by_type": {
"choice": {
"correct": 223,
"n": 600,
"accuracy": 0.37166666666666665
},
"noul": {
"correct": 296,
"n": 600,
"accuracy": 0.49333333333333335
},
"score": {
"correct": 359,
"n": 800,
"accuracy": 0.44875
}
},
"ci95_cases": [
0.4185,
0.4595
],
"score_mae": {
"mae_vs_gold_expectation": 0.5391987730165998,
"mae_vs_gold_label": 0.7372078359872103,
"n": 800
}
},
"v3|telepatia_es|full-case joint": {
"n": 2000,
"correct": 745,
"accuracy": 0.3725,
"nll": 1.1746715414243962,
"kl": 0.42382525863971043,
"ce_soft": 1.1911838142640392,
"brier": 0.235579264029217,
"brier_mean_k": 0.07444738062797601,
"brier_onehot": 0.6762935895205119,
"ece10": 0.08313308775424957,
"mean_conf": 0.4297241877913475,
"by_type": {
"choice": {
"correct": 230,
"n": 600,
"accuracy": 0.38333333333333336
},
"noul": {
"correct": 279,
"n": 600,
"accuracy": 0.465
},
"score": {
"correct": 236,
"n": 800,
"accuracy": 0.295
}
},
"ci95_cases": [
0.35,
0.3945
],
"score_mae": {
"mae_vs_gold_expectation": 0.6721788485455155,
"mae_vs_gold_label": 0.8984525854373351,
"n": 800
}
},
"v3|telepatia_es|single-question vs full-case context": {
"n_common": 2000,
"only_b_right": 136,
"only_a_right": 784,
"p_mcnemar_exact": 2.4222306432783787e-111
},
"v3|telepatia_es|full-case context vs full-case joint": {
"n_common": 2000,
"only_b_right": 330,
"only_a_right": 463,
"p_mcnemar_exact": 2.6285909419854747e-06
}
}