experiments / swap_analysis /results_mcq /molmo /json /cross_alignment_2m.json
ch-min's picture
Add files using upload-large-folder tool
9a8f4de verified
raw
history blame
8.46 kB
{
"L0": {
"per_sample_mean": -0.009254118177338497,
"per_sample_std": 0.06464908030731169,
"mean_delta_alignment": -0.009769584983587265,
"permutation_mean": -1.9074110241490416e-05,
"permutation_std": 0.001976528437808156,
"n_samples": 1039
},
"L1": {
"per_sample_mean": 0.11192101686012516,
"per_sample_std": 0.1693834856912801,
"mean_delta_alignment": 0.04189475253224373,
"permutation_mean": 0.00024084997130557895,
"permutation_std": 0.003457510843873024,
"n_samples": 1039
},
"L2": {
"per_sample_mean": 0.14013125071185753,
"per_sample_std": 0.13558443809621898,
"mean_delta_alignment": 0.1215815395116806,
"permutation_mean": 2.2322645236272365e-05,
"permutation_std": 0.0025323566514998674,
"n_samples": 1039
},
"L3": {
"per_sample_mean": 0.052659211419360016,
"per_sample_std": 0.12332778584226863,
"mean_delta_alignment": 0.13099607825279236,
"permutation_mean": 0.00023801330826245248,
"permutation_std": 0.0021408838219940662,
"n_samples": 1039
},
"L4": {
"per_sample_mean": 0.1047383931667437,
"per_sample_std": 0.11801041321247449,
"mean_delta_alignment": 0.20510317385196686,
"permutation_mean": 0.00012114084529457614,
"permutation_std": 0.002394289243966341,
"n_samples": 1039
},
"L5": {
"per_sample_mean": 0.12467854293782685,
"per_sample_std": 0.10003065293569505,
"mean_delta_alignment": 0.21715739369392395,
"permutation_mean": 8.518288086634129e-05,
"permutation_std": 0.00224831304512918,
"n_samples": 1039
},
"L6": {
"per_sample_mean": 0.10576482900402431,
"per_sample_std": 0.08687350192741541,
"mean_delta_alignment": 0.1667506843805313,
"permutation_mean": -8.926590089686215e-05,
"permutation_std": 0.002062375657260418,
"n_samples": 1039
},
"L7": {
"per_sample_mean": 0.10800409257375458,
"per_sample_std": 0.07344973691784622,
"mean_delta_alignment": 0.18538077175617218,
"permutation_mean": -0.00014648574870079756,
"permutation_std": 0.0018356518121436238,
"n_samples": 1039
},
"L8": {
"per_sample_mean": 0.10751237613617337,
"per_sample_std": 0.06935517593491461,
"mean_delta_alignment": 0.1713843047618866,
"permutation_mean": -0.00021166319493204355,
"permutation_std": 0.0016664436552673578,
"n_samples": 1039
},
"L9": {
"per_sample_mean": 0.13032337937540917,
"per_sample_std": 0.09954265645706266,
"mean_delta_alignment": 0.13535599410533905,
"permutation_mean": -0.00010368316725362092,
"permutation_std": 0.0029568492900580168,
"n_samples": 1039
},
"L10": {
"per_sample_mean": 0.1274741151499885,
"per_sample_std": 0.10990764249346725,
"mean_delta_alignment": -0.028085293248295784,
"permutation_mean": -0.00026070530293509364,
"permutation_std": 0.002894388046115637,
"n_samples": 1039
},
"L11": {
"per_sample_mean": 0.09229658159058968,
"per_sample_std": 0.09848588137104661,
"mean_delta_alignment": 0.028461644425988197,
"permutation_mean": -1.2980699466424994e-05,
"permutation_std": 0.002497854409739375,
"n_samples": 1039
},
"L12": {
"per_sample_mean": 0.1965631897100409,
"per_sample_std": 0.11088623783832603,
"mean_delta_alignment": 0.21971772611141205,
"permutation_mean": 0.00032745927455835044,
"permutation_std": 0.002732295310124755,
"n_samples": 1039
},
"L13": {
"per_sample_mean": 0.3969216635773412,
"per_sample_std": 0.1435224044379739,
"mean_delta_alignment": 0.48838749527931213,
"permutation_mean": 0.0014766021631658077,
"permutation_std": 0.006168306805193424,
"n_samples": 1039
},
"L14": {
"per_sample_mean": 0.7382648016708877,
"per_sample_std": 0.0710094818172151,
"mean_delta_alignment": 0.8166166543960571,
"permutation_mean": 0.0016780676087364554,
"permutation_std": 0.008902635425329208,
"n_samples": 1039
},
"L15": {
"per_sample_mean": 0.7359273408443682,
"per_sample_std": 0.0639730397117569,
"mean_delta_alignment": 0.8333110213279724,
"permutation_mean": 0.0018688386771827936,
"permutation_std": 0.008581453934311867,
"n_samples": 1039
},
"L16": {
"per_sample_mean": 0.733613237093227,
"per_sample_std": 0.0716905782156783,
"mean_delta_alignment": 0.8327218294143677,
"permutation_mean": 0.0020154076628386974,
"permutation_std": 0.00829540565609932,
"n_samples": 1039
},
"L17": {
"per_sample_mean": 0.7850513141584351,
"per_sample_std": 0.0662623843274579,
"mean_delta_alignment": 0.8615885972976685,
"permutation_mean": 0.0018261586083099246,
"permutation_std": 0.009291241876780987,
"n_samples": 1039
},
"L18": {
"per_sample_mean": 0.5450107111776861,
"per_sample_std": 0.2568719286920903,
"mean_delta_alignment": 0.8521422743797302,
"permutation_mean": 0.002654711715877056,
"permutation_std": 0.009685528464615345,
"n_samples": 1039
},
"L19": {
"per_sample_mean": 0.41777046441146587,
"per_sample_std": 0.402112155531179,
"mean_delta_alignment": 0.8543694019317627,
"permutation_mean": 0.0033813754562288523,
"permutation_std": 0.01249579805880785,
"n_samples": 1039
},
"L20": {
"per_sample_mean": 0.4110572521288351,
"per_sample_std": 0.390758582344652,
"mean_delta_alignment": 0.8421671986579895,
"permutation_mean": 0.00323235965333879,
"permutation_std": 0.012163028120994568,
"n_samples": 1039
},
"L21": {
"per_sample_mean": 0.3367718115264749,
"per_sample_std": 0.48547811324017726,
"mean_delta_alignment": 0.8689522743225098,
"permutation_mean": 0.0038173028733581305,
"permutation_std": 0.014513869769871235,
"n_samples": 1039
},
"L22": {
"per_sample_mean": 0.2880282264364851,
"per_sample_std": 0.536806263893991,
"mean_delta_alignment": 0.8777095675468445,
"permutation_mean": 0.004165720194578171,
"permutation_std": 0.015951497480273247,
"n_samples": 1039
},
"L23": {
"per_sample_mean": 0.28185969071787237,
"per_sample_std": 0.5369275939496753,
"mean_delta_alignment": 0.8632226586341858,
"permutation_mean": 0.0041371798142790794,
"permutation_std": 0.015933474525809288,
"n_samples": 1039
},
"L24": {
"per_sample_mean": 0.283797710602805,
"per_sample_std": 0.5301618955075025,
"mean_delta_alignment": 0.8572227358818054,
"permutation_mean": 0.003987807314842939,
"permutation_std": 0.015684397891163826,
"n_samples": 1039
},
"L25": {
"per_sample_mean": 0.2806209442956432,
"per_sample_std": 0.5308109019242917,
"mean_delta_alignment": 0.8469801545143127,
"permutation_mean": 0.003983200993388891,
"permutation_std": 0.01563182659447193,
"n_samples": 1039
},
"L26": {
"per_sample_mean": 0.2743802673158844,
"per_sample_std": 0.5329466471993541,
"mean_delta_alignment": 0.8442679643630981,
"permutation_mean": 0.003967928700149059,
"permutation_std": 0.015648072585463524,
"n_samples": 1039
},
"L27": {
"per_sample_mean": 0.26589061796290514,
"per_sample_std": 0.5345333850716167,
"mean_delta_alignment": 0.8364192247390747,
"permutation_mean": 0.003892181208357215,
"permutation_std": 0.01567050628364086,
"n_samples": 1039
},
"L28": {
"per_sample_mean": 0.2660821527112688,
"per_sample_std": 0.5201851314632757,
"mean_delta_alignment": 0.8134483098983765,
"permutation_mean": 0.003744215937331319,
"permutation_std": 0.015345877967774868,
"n_samples": 1039
},
"L29": {
"per_sample_mean": 0.22459347781844444,
"per_sample_std": 0.5438565183193912,
"mean_delta_alignment": 0.8072983622550964,
"permutation_mean": 0.0038039570208638906,
"permutation_std": 0.016142431646585464,
"n_samples": 1039
},
"L30": {
"per_sample_mean": 0.23420973230221445,
"per_sample_std": 0.5035483634127362,
"mean_delta_alignment": 0.7649968266487122,
"permutation_mean": 0.0033226595260202885,
"permutation_std": 0.015032452531158924,
"n_samples": 1039
},
"L31": {
"per_sample_mean": 0.23679784329180517,
"per_sample_std": 0.47009771386555604,
"mean_delta_alignment": 0.6918860077857971,
"permutation_mean": 0.0028441923204809427,
"permutation_std": 0.014296585693955421,
"n_samples": 1039
}
}