GLM-5.3-speculator.dflash2-preview / data /verified_evaluation_rows.json
shanjiaz's picture
Add verified DFlash2 evaluation model card and graphs
54b9231 verified
Raw History Blame Contribute Delete
15.3 kB
[
{
"experiment": "BASIC4 T0",
"dataset": "HumanEval",
"successful_calls": 6113,
"errors": 0,
"acceptance_length": 4.807433568954425,
"acceptance_rate_pct": 47.59291961193031,
"position_1_pct": 83.48910457231432,
"position_2_pct": 68.2955461615601,
"position_3_pct": 56.05040420414842,
"position_4_pct": 46.438411701067665,
"position_5_pct": 39.082074736092274,
"position_6_pct": 33.35858251014397,
"position_7_pct": 28.844422877439253,
"position_8_pct": 25.184810132676535
},
{
"experiment": "BASIC4 T0",
"dataset": "GSM8K",
"successful_calls": 112565,
"errors": 0,
"acceptance_length": 5.8452949929301266,
"acceptance_rate_pct": 60.566187411626586,
"position_1_pct": 89.111124649917,
"position_2_pct": 78.43314942568261,
"position_3_pct": 69.42534995217603,
"position_4_pct": 61.56513870444342,
"position_5_pct": 54.91736490150955,
"position_6_pct": 48.928548868838995,
"position_7_pct": 43.53965352202276,
"position_8_pct": 38.60916926842232
},
{
"experiment": "BASIC4 T0",
"dataset": "MATH500",
"successful_calls": 17996,
"errors": 0,
"acceptance_length": 5.315660395956892,
"acceptance_rate_pct": 53.945754949461154,
"position_1_pct": 86.4338302641223,
"position_2_pct": 73.8404691747072,
"position_3_pct": 63.20606152624376,
"position_4_pct": 54.37304319945573,
"position_5_pct": 47.034425086389675,
"position_6_pct": 40.77491824898309,
"position_7_pct": 35.334843618094816,
"position_8_pct": 30.568448477692662
},
{
"experiment": "BASIC4 T0",
"dataset": "MT-Bench",
"successful_calls": 4875,
"errors": 0,
"acceptance_length": 3.6381420740851347,
"acceptance_rate_pct": 32.976775926064185,
"position_1_pct": 74.9640303021086,
"position_2_pct": 54.310188295953424,
"position_3_pct": 39.48269998092062,
"position_4_pct": 29.28122358727341,
"position_5_pct": 22.363451100105905,
"position_6_pct": 17.565171746666408,
"position_7_pct": 14.180632863906885,
"position_8_pct": 11.666809531578215
},
{
"experiment": "SPEED-Bench T0",
"dataset": "coding",
"successful_calls": 89,
"errors": 0,
"acceptance_length": 4.0971,
"acceptance_rate_pct": 38.71,
"position_1_pct": 78.33,
"position_2_pct": 60.15,
"position_3_pct": 46.26,
"position_4_pct": 36.22,
"position_5_pct": 28.99,
"position_6_pct": 23.66,
"position_7_pct": 19.59,
"position_8_pct": 16.51
},
{
"experiment": "SPEED-Bench T0",
"dataset": "humanities",
"successful_calls": 88,
"errors": 0,
"acceptance_length": 3.8826,
"acceptance_rate_pct": 36.03,
"position_1_pct": 76.53,
"position_2_pct": 57.73,
"position_3_pct": 43.58,
"position_4_pct": 33.12,
"position_5_pct": 25.87,
"position_6_pct": 20.46,
"position_7_pct": 16.82,
"position_8_pct": 14.16
},
{
"experiment": "SPEED-Bench T0",
"dataset": "math",
"successful_calls": 98,
"errors": 0,
"acceptance_length": 3.9028,
"acceptance_rate_pct": 36.28,
"position_1_pct": 77.84,
"position_2_pct": 59.14,
"position_3_pct": 44.81,
"position_4_pct": 34.08,
"position_5_pct": 26.16,
"position_6_pct": 20.21,
"position_7_pct": 15.73,
"position_8_pct": 12.3
},
{
"experiment": "SPEED-Bench T0",
"dataset": "qa",
"successful_calls": 80,
"errors": 0,
"acceptance_length": 4.2508,
"acceptance_rate_pct": 40.64,
"position_1_pct": 80.65,
"position_2_pct": 63.71,
"position_3_pct": 49.73,
"position_4_pct": 38.84,
"position_5_pct": 30.85,
"position_6_pct": 24.68,
"position_7_pct": 20.06,
"position_8_pct": 16.56
},
{
"experiment": "SPEED-Bench T0",
"dataset": "rag",
"successful_calls": 105,
"errors": 0,
"acceptance_length": 4.9694,
"acceptance_rate_pct": 49.62,
"position_1_pct": 84.57,
"position_2_pct": 70.38,
"position_3_pct": 58.39,
"position_4_pct": 48.39,
"position_5_pct": 41.03,
"position_6_pct": 35.35,
"position_7_pct": 31.23,
"position_8_pct": 27.59
},
{
"experiment": "SPEED-Bench T0",
"dataset": "reasoning",
"successful_calls": 190,
"errors": 0,
"acceptance_length": 4.5232,
"acceptance_rate_pct": 44.04,
"position_1_pct": 81.24,
"position_2_pct": 65.24,
"position_3_pct": 51.8,
"position_4_pct": 41.39,
"position_5_pct": 34.17,
"position_6_pct": 29.37,
"position_7_pct": 25.84,
"position_8_pct": 23.27
},
{
"experiment": "SPEED-Bench T0",
"dataset": "stem",
"successful_calls": 86,
"errors": 0,
"acceptance_length": 3.8932,
"acceptance_rate_pct": 36.17,
"position_1_pct": 77.98,
"position_2_pct": 59.32,
"position_3_pct": 44.71,
"position_4_pct": 33.55,
"position_5_pct": 25.57,
"position_6_pct": 19.78,
"position_7_pct": 15.73,
"position_8_pct": 12.67
},
{
"experiment": "SPEED-Bench T0",
"dataset": "writing",
"successful_calls": 84,
"errors": 0,
"acceptance_length": 3.1636,
"acceptance_rate_pct": 27.04,
"position_1_pct": 70.81,
"position_2_pct": 48.53,
"position_3_pct": 32.9,
"position_4_pct": 22.57,
"position_5_pct": 15.76,
"position_6_pct": 11.23,
"position_7_pct": 8.27,
"position_8_pct": 6.27
},
{
"experiment": "SPEED-Bench T0",
"dataset": "multilingual",
"successful_calls": 80,
"errors": 0,
"acceptance_length": 3.9722,
"acceptance_rate_pct": 37.15,
"position_1_pct": 75.38,
"position_2_pct": 57.57,
"position_3_pct": 44.4,
"position_4_pct": 35.04,
"position_5_pct": 28.12,
"position_6_pct": 22.69,
"position_7_pct": 18.64,
"position_8_pct": 15.38
},
{
"experiment": "SPEED-Bench T0",
"dataset": "summarization",
"successful_calls": 80,
"errors": 0,
"acceptance_length": 4.629,
"acceptance_rate_pct": 45.36,
"position_1_pct": 85.01,
"position_2_pct": 69.37,
"position_3_pct": 55.91,
"position_4_pct": 45.3,
"position_5_pct": 36.43,
"position_6_pct": 28.88,
"position_7_pct": 23.32,
"position_8_pct": 18.67
},
{
"experiment": "SPEED-Bench T0",
"dataset": "roleplay",
"successful_calls": 158,
"errors": 0,
"acceptance_length": 2.9572,
"acceptance_rate_pct": 24.47,
"position_1_pct": 67.65,
"position_2_pct": 43.17,
"position_3_pct": 27.69,
"position_4_pct": 18.57,
"position_5_pct": 13.21,
"position_6_pct": 10.11,
"position_7_pct": 8.3,
"position_8_pct": 7.01
},
{
"experiment": "BASIC4 T1",
"dataset": "humaneval",
"successful_calls": 164,
"errors": 0,
"acceptance_length": 4.294067058567386,
"acceptance_rate_pct": 41.17583823209232,
"position_1_pct": 78.17929457870673,
"position_2_pct": 61.04670150228609,
"position_3_pct": 48.516220335292836,
"position_4_pct": 39.07794469845417,
"position_5_pct": 32.30677117352493,
"position_6_pct": 27.15708687132593,
"position_7_pct": 23.161332462442846,
"position_8_pct": 19.961354234704984
},
{
"experiment": "BASIC4 T1",
"dataset": "gsm8k",
"successful_calls": 1319,
"errors": 0,
"acceptance_length": 4.662285405179695,
"acceptance_rate_pct": 45.77856756474619,
"position_1_pct": 79.8866882408207,
"position_2_pct": 64.5829576836571,
"position_3_pct": 53.301130270562666,
"position_4_pct": 44.72777656515426,
"position_5_pct": 37.97841951923168,
"position_6_pct": 32.70572157953176,
"position_7_pct": 28.346919988991488,
"position_8_pct": 24.698926670019834
},
{
"experiment": "BASIC4 T1",
"dataset": "math500",
"successful_calls": 500,
"errors": 0,
"acceptance_length": 4.632882069795427,
"acceptance_rate_pct": 45.41102587244284,
"position_1_pct": 80.61321700762134,
"position_2_pct": 65.35148415563577,
"position_3_pct": 53.61311672683514,
"position_4_pct": 44.62845968712395,
"position_5_pct": 37.59025270758123,
"position_6_pct": 31.784997994384277,
"position_7_pct": 26.94494584837545,
"position_8_pct": 22.76173285198556
},
{
"experiment": "BASIC4 T1",
"dataset": "mtbench",
"successful_calls": 160,
"errors": 0,
"acceptance_length": 3.095874052030335,
"acceptance_rate_pct": 26.19842565037919,
"position_1_pct": 67.74426418354612,
"position_2_pct": 45.31208601324757,
"position_3_pct": 30.91177882307766,
"position_4_pct": 21.625803974272824,
"position_5_pct": 15.704329461457233,
"position_6_pct": 11.828741480272631,
"position_7_pct": 9.20840933090141,
"position_8_pct": 7.251991936258039
},
{
"experiment": "SPEED-Bench T1",
"dataset": "coding",
"successful_calls": 89,
"errors": 0,
"acceptance_length": 3.7522630712699994,
"acceptance_rate_pct": 34.40328839087499,
"position_1_pct": 73.83975732986298,
"position_2_pct": 54.35868483503024,
"position_3_pct": 40.78217101737732,
"position_4_pct": 31.3069356196892,
"position_5_pct": 24.800486105794995,
"position_6_pct": 19.984402510908673,
"position_7_pct": 16.45774324427773,
"position_8_pct": 13.696126464058793
},
{
"experiment": "SPEED-Bench T1",
"dataset": "humanities",
"successful_calls": 78,
"errors": 10,
"acceptance_length": 2.8185892204995118,
"acceptance_rate_pct": 22.732365256243895,
"position_1_pct": 63.61483820047356,
"position_2_pct": 40.73883322408465,
"position_3_pct": 26.604506396621762,
"position_4_pct": 17.83451277317744,
"position_5_pct": 12.423804618766694,
"position_6_pct": 8.951034295167641,
"position_7_pct": 6.6441034696489325,
"position_8_pct": 5.0472890720104875
},
{
"experiment": "SPEED-Bench T1",
"dataset": "math",
"successful_calls": 88,
"errors": 10,
"acceptance_length": 3.481256654324715,
"acceptance_rate_pct": 31.01570817905894,
"position_1_pct": 72.07726167585692,
"position_2_pct": 51.40952502026093,
"position_3_pct": 37.66519940462002,
"position_4_pct": 27.85493704545334,
"position_5_pct": 21.03031458733917,
"position_6_pct": 16.058044250928294,
"position_7_pct": 12.416638857548454,
"position_8_pct": 9.613744590464385
},
{
"experiment": "SPEED-Bench T1",
"dataset": "multilingual",
"successful_calls": 80,
"errors": 0,
"acceptance_length": 3.498545566544608,
"acceptance_rate_pct": 31.231819581807603,
"position_1_pct": 69.5424523459156,
"position_2_pct": 50.07699941822662,
"position_3_pct": 37.05383114883132,
"position_4_pct": 28.233120016426543,
"position_5_pct": 21.867834776359466,
"position_6_pct": 17.454912562882857,
"position_7_pct": 14.140515382772664,
"position_8_pct": 11.484891003045755
},
{
"experiment": "SPEED-Bench T1",
"dataset": "qa",
"successful_calls": 79,
"errors": 1,
"acceptance_length": 3.1040785289644686,
"acceptance_rate_pct": 26.300981612055857,
"position_1_pct": 69.17185123738422,
"position_2_pct": 47.02889534079912,
"position_3_pct": 32.15816397068989,
"position_4_pct": 22.065532973869765,
"position_5_pct": 15.285497027512788,
"position_6_pct": 10.996820129959906,
"position_7_pct": 7.902668325729296,
"position_8_pct": 5.798423890501867
},
{
"experiment": "SPEED-Bench T1",
"dataset": "rag",
"successful_calls": 105,
"errors": 0,
"acceptance_length": 3.8383407546285446,
"acceptance_rate_pct": 35.47925943285681,
"position_1_pct": 72.8521209280525,
"position_2_pct": 54.408249355519104,
"position_3_pct": 42.06233887977502,
"position_4_pct": 33.016170611670965,
"position_5_pct": 26.833841106163582,
"position_6_pct": 21.837356456526834,
"position_7_pct": 17.956409655495666,
"position_8_pct": 14.867588469650808
},
{
"experiment": "SPEED-Bench T1",
"dataset": "reasoning",
"successful_calls": 189,
"errors": 1,
"acceptance_length": 2.8580765924236786,
"acceptance_rate_pct": 23.225957405295986,
"position_1_pct": 63.283560249007685,
"position_2_pct": 40.626543341605135,
"position_3_pct": 26.90917072635372,
"position_4_pct": 18.538390717695766,
"position_5_pct": 13.239965612062605,
"position_6_pct": 9.87189883605368,
"position_7_pct": 7.515349580211083,
"position_8_pct": 5.822780179378213
},
{
"experiment": "SPEED-Bench T1",
"dataset": "roleplay",
"successful_calls": 158,
"errors": 0,
"acceptance_length": 2.4504565641942078,
"acceptance_rate_pct": 18.130707052427596,
"position_1_pct": 61.07780819025645,
"position_2_pct": 35.241779205404875,
"position_3_pct": 20.091840658719207,
"position_4_pct": 11.655016927937506,
"position_5_pct": 7.141402945234917,
"position_6_pct": 4.555839572918316,
"position_7_pct": 3.0832221141448186,
"position_8_pct": 2.198746804804669
},
{
"experiment": "SPEED-Bench T1",
"dataset": "stem",
"successful_calls": 78,
"errors": 8,
"acceptance_length": 3.002715318869165,
"acceptance_rate_pct": 25.033941485864563,
"position_1_pct": 66.85820731974579,
"position_2_pct": 44.85163269778654,
"position_3_pct": 30.34845496383958,
"position_4_pct": 20.8531667762437,
"position_5_pct": 14.379793995178611,
"position_6_pct": 10.174008327854482,
"position_7_pct": 7.408941485864562,
"position_8_pct": 5.397326320403243
},
{
"experiment": "SPEED-Bench T1",
"dataset": "summarization",
"successful_calls": 80,
"errors": 0,
"acceptance_length": 4.29741136747327,
"acceptance_rate_pct": 41.21764209341587,
"position_1_pct": 81.46689176514725,
"position_2_pct": 64.72519227161884,
"position_3_pct": 51.15362971299944,
"position_4_pct": 40.20821609454136,
"position_5_pct": 31.523166385293568,
"position_6_pct": 24.826486587882197,
"position_7_pct": 19.864940911648848,
"position_8_pct": 15.97261301819546
},
{
"experiment": "SPEED-Bench T1",
"dataset": "writing",
"successful_calls": 80,
"errors": 4,
"acceptance_length": 2.772608882265594,
"acceptance_rate_pct": 22.15761102831992,
"position_1_pct": 64.7225991517087,
"position_2_pct": 41.07388025283453,
"position_3_pct": 26.214386828291662,
"position_4_pct": 16.991685039202146,
"position_5_pct": 11.295443700064249,
"position_6_pct": 7.687406526671056,
"position_7_pct": 5.4181873438778805,
"position_8_pct": 3.8572993839091247
},
{
"experiment": "MRCR T0",
"dataset": "8 needles",
"successful_calls": 458,
"errors": 342,
"acceptance_length": 6.207113290053785,
"acceptance_rate_pct": 65.08891612567231,
"position_1_pct": 92.19037613698606,
"position_2_pct": 84.60747604809238,
"position_3_pct": 75.13221812173944,
"position_4_pct": 65.37950417081953,
"position_5_pct": 57.71399581571173,
"position_6_pct": 52.20124091982509,
"position_7_pct": 48.23020768795625,
"position_8_pct": 45.25631010424804
}
]