{ "456": { "bp": { "log": { "train_loss": [ 1.932577501449585, 1.6900249602508546, 1.6011367318725587, 1.5466754390716553, 1.4920579125213622, 1.4546358205413819, 1.4231391117858887, 1.3938725402450562, 1.3629425603866576, 1.342505719642639, 1.3237205046844482, 1.300122989501953, 1.2825427363967896, 1.2623162283325196, 1.2426594482040405, 1.227365916481018, 1.206711524925232, 1.190613479347229, 1.1721268761444092, 1.1622920285797118, 1.1426683759689331, 1.1272672763824463, 1.1123097320747375, 1.095245484790802, 1.0796047261238098, 1.0664704471588136, 1.0544639612579345, 1.0334472966003418, 1.0216301371002197, 1.0061453499603272, 0.9892120691299439, 0.9690513248062134, 0.9597649993705749, 0.9393893640518188, 0.9262461110305786, 0.9142684574699402, 0.8982764770317078, 0.882060844745636, 0.8589154948616028, 0.8484143830871582, 0.8298149911689758, 0.8117531860733033, 0.7923431532669067, 0.7819426927185058, 0.7630360532951355, 0.7432640241622925, 0.725852869091034, 0.7111906971359253, 0.6900048795890809, 0.6697983954048157, 0.6541705364608764, 0.6370760979270935, 0.6198028992843628, 0.5972342865371704, 0.5801713646697998, 0.5637391392326355, 0.5492425328063965, 0.5351220473480225, 0.517744523601532, 0.4910259468269348, 0.48044670740127565, 0.4666313941383362, 0.4454871296596527, 0.4287984812068939, 0.4167753814029694, 0.4019090121555328, 0.3892905419158936, 0.3745353739929199, 0.3589991714382172, 0.34674671503067017, 0.3343308783531189, 0.32089566977500916, 0.3101300811481476, 0.2954042288208008, 0.2848380702114105, 0.2828438286781311, 0.264346894235611, 0.26094838644981383, 0.25063199663162233, 0.24315124837875365, 0.2345473566555977, 0.22507079369306565, 0.2194176360464096, 0.21162880234241485, 0.20445122440338134, 0.2012126348876953, 0.19840873272895812, 0.19022230600357057, 0.1889617627120018, 0.18509810159683227, 0.18254041098594664, 0.1836097541332245, 0.17741597022533417, 0.1746750330734253, 0.17486595999717713, 0.17480743077754973, 0.17309590301036834, 0.17227657488822937, 0.1740615609550476, 0.16915745746612548 ], "train_acc": [ 0.30628, 0.38592, 0.42152, 0.4394, 0.45908, 0.47766, 0.48482, 0.49718, 0.51004, 0.51544, 0.52028, 0.53156, 0.53764, 0.54346, 0.55342, 0.55874, 0.56678, 0.571, 0.57738, 0.57968, 0.59098, 0.59448, 0.59972, 0.60702, 0.61296, 0.61686, 0.62066, 0.629, 0.63218, 0.63706, 0.6458, 0.65254, 0.6557, 0.66064, 0.66696, 0.67284, 0.67792, 0.68442, 0.6895, 0.69484, 0.69998, 0.70706, 0.71588, 0.71756, 0.72552, 0.7343, 0.74024, 0.74394, 0.75236, 0.76022, 0.76464, 0.77054, 0.7777, 0.78372, 0.79538, 0.7972, 0.80364, 0.8092, 0.81504, 0.82654, 0.82918, 0.83316, 0.84092, 0.84854, 0.85006, 0.85778, 0.8616, 0.8676, 0.87464, 0.87672, 0.88258, 0.88796, 0.89154, 0.89712, 0.90188, 0.90184, 0.90858, 0.90758, 0.9118, 0.91542, 0.9187, 0.92286, 0.9233, 0.92722, 0.93012, 0.93172, 0.93268, 0.936, 0.93634, 0.9374, 0.93696, 0.93802, 0.94054, 0.94164, 0.9409, 0.9415, 0.94154, 0.94296, 0.94158, 0.94336 ], "test_acc": [ 0.3915, 0.4289, 0.4595, 0.4775, 0.4915, 0.4936, 0.5132, 0.5139, 0.5291, 0.5317, 0.5407, 0.5449, 0.5481, 0.5499, 0.5545, 0.5589, 0.5696, 0.5674, 0.5737, 0.5776, 0.573, 0.583, 0.5773, 0.572, 0.5836, 0.5774, 0.5838, 0.5796, 0.5884, 0.5942, 0.5925, 0.5927, 0.5974, 0.5934, 0.5962, 0.5972, 0.5995, 0.595, 0.5921, 0.5848, 0.5907, 0.5934, 0.5935, 0.5921, 0.5969, 0.5936, 0.5961, 0.5948, 0.5992, 0.5954, 0.5978, 0.5933, 0.5934, 0.5987, 0.5987, 0.591, 0.5889, 0.5924, 0.5937, 0.593, 0.5925, 0.5944, 0.5938, 0.5924, 0.5946, 0.5905, 0.5928, 0.5899, 0.5891, 0.587, 0.5891, 0.5846, 0.5867, 0.586, 0.5865, 0.585, 0.5851, 0.5824, 0.5821, 0.5855, 0.5825, 0.5861, 0.5857, 0.5869, 0.5855, 0.589, 0.5887, 0.5872, 0.5863, 0.5857, 0.5865, 0.5868, 0.5867, 0.5874, 0.5872, 0.5876, 0.5869, 0.5879, 0.5871, 0.587 ] }, "diagnostics": { "bp_cosine": [ 0.9672282934188843, 0.9684309363365173, 0.9675205945968628, 0.9669639468193054, 0.9665698409080505, 0.967890739440918, 0.9664402008056641, 0.9660602807998657, 0.9652657508850098, 0.9614400863647461, 0.9602784514427185, 0.9578725099563599 ], "perturbation_rho": [ 0.9848342537879944, 0.9891709685325623, 0.9909648895263672, 0.9893825054168701, 0.9924193024635315, 0.9927305579185486, 0.9941403865814209, 0.9934987425804138, 0.9928717017173767, 0.9915444254875183, 0.9898866415023804, 0.9873980283737183 ], "nudging": { "0.001": [ -0.0028492475394159555, -0.0028252778574824333, -0.0028643091209232807, -0.0029328675009310246, -0.002945749554783106, -0.0029587389435619116, -0.00288290623575449, -0.0028087471146136522, -0.0026325639337301254, -0.002420567674562335, -0.002228461904451251, -0.0018789597088471055 ], "0.003": [ -0.008540419861674309, -0.008468939922749996, -0.008585674688220024, -0.008790891617536545, -0.008829280734062195, -0.008868295699357986, -0.00864122249186039, -0.008419105783104897, -0.007891927845776081, -0.007256433367729187, -0.006680862978100777, -0.0056340815499424934 ], "0.01": [ -0.028379585593938828, -0.028143716976046562, -0.028530558571219444, -0.02920939214527607, -0.02933822199702263, -0.029468875378370285, -0.028714874759316444, -0.027982797473669052, -0.026236172765493393, -0.024128224700689316, -0.022218093276023865, -0.018745578825473785 ] } }, "drift": { "embed.weight": 8.834297843591171, "embed.bias": 15.413548194165312, "blocks.0.ln.weight": 0.3280907780755125, "blocks.0.w1.weight": 4.220952328018163, "blocks.0.w1.bias": 6.7337611045609105, "blocks.0.w2.weight": 10.671510043086428, "blocks.1.ln.weight": 0.22467724575060716, "blocks.1.w1.weight": 4.523437520518719, "blocks.1.w1.bias": 4.4537544678846235, "blocks.1.w2.weight": 11.922866714123732, "blocks.2.ln.weight": 0.19572857698834295, "blocks.2.w1.weight": 4.598553685356083, "blocks.2.w1.bias": 4.125724114621679, "blocks.2.w2.weight": 12.310215404426595, "blocks.3.ln.weight": 0.21390126577946425, "blocks.3.w1.weight": 4.584201339373962, "blocks.3.w1.bias": 3.7580021741696004, "blocks.3.w2.weight": 12.601184193203743, "blocks.4.ln.weight": 0.24161530581374482, "blocks.4.w1.weight": 4.518574188984943, "blocks.4.w1.bias": 3.8030594737653343, "blocks.4.w2.weight": 12.669272886895255, "blocks.5.ln.weight": 0.25422945739387653, "blocks.5.w1.weight": 4.461826473966918, "blocks.5.w1.bias": 3.6451227277370744, "blocks.5.w2.weight": 12.680759003338304, "blocks.6.ln.weight": 0.25745358599301726, "blocks.6.w1.weight": 4.47036651988774, "blocks.6.w1.bias": 3.706163791416811, "blocks.6.w2.weight": 13.119190581706045, "blocks.7.ln.weight": 0.263116520949555, "blocks.7.w1.weight": 4.434269376799208, "blocks.7.w1.bias": 3.6960182135036708, "blocks.7.w2.weight": 13.098144437096936, "blocks.8.ln.weight": 0.27514230251440513, "blocks.8.w1.weight": 4.344487791539744, "blocks.8.w1.bias": 3.648654542354885, "blocks.8.w2.weight": 13.035629633632155, "blocks.9.ln.weight": 0.29172129249550904, "blocks.9.w1.weight": 4.255456615430207, "blocks.9.w1.bias": 3.6968063474376733, "blocks.9.w2.weight": 13.086873727145367, "blocks.10.ln.weight": 0.3000076159356563, "blocks.10.w1.weight": 4.2734266858824945, "blocks.10.w1.bias": 3.6213422959644603, "blocks.10.w2.weight": 13.200754189456744, "blocks.11.ln.weight": 0.2902311063472059, "blocks.11.w1.weight": 4.144956242341598, "blocks.11.w1.bias": 3.5713523930610735, "blocks.11.w2.weight": 13.254643701043216, "out_ln.weight": 0.3155177407468068, "out_head.weight": 3.523856665191762, "out_head.bias": 1.0793116101743525 } }, "dfa": { "log": { "train_loss": [ 2.07058688041687, 2.0397231711578367, 2.040001149024963, 2.036939584312439, 2.035007101402283, 2.0326563871002197, 2.0309880726623537, 2.028131689300537, 2.0242818844223023, 2.021325262145996, 2.01944966506958, 2.0185741081237794, 2.0193587062072753, 2.0140840717315673, 2.01566628868103, 2.0162557276153565, 2.01352592338562, 2.0122515713882447, 2.007948448524475, 2.0104500115966797, 2.008844228897095, 2.007831936798096, 2.0087617783355713, 2.0067150954437256, 2.0060685678100585, 2.0067182321548462, 2.004691079711914, 2.004318997192383, 2.006708196105957, 2.0059428287506105, 2.002060938720703, 2.0030168786239626, 2.0013856718444822, 2.001412846107483, 2.0029963275909424, 2.001156611251831, 2.0020238697052, 2.0013174156951905, 2.001038536148071, 1.9974690299224853, 1.9983380174255372, 2.00068127532959, 2.002073029937744, 1.9994858475494384, 1.999331954574585, 1.9974100791168212, 1.9981875708389283, 1.9986576483917236, 1.9978187020492553, 1.9974584699249267, 1.9957377856063843, 1.995005799217224, 1.9989547797012328, 1.997181711654663, 1.995425957107544, 1.9960723597717285, 1.996761396522522, 1.9958302701187134, 1.9968557986450195, 1.9969658164215087, 1.9949297325134276, 1.9950204270172118, 1.9955382054901123, 1.9956761499786377, 1.9941519859313965, 1.9952024503326415, 1.995510894241333, 1.9926798336791993, 1.9947078677749635, 1.9926326573181152, 1.9923986828231812, 1.993803144607544, 1.9924064002227784, 1.9935845022583008, 1.993097350692749, 1.9917376801300049, 1.9939163592147826, 1.9913390029144287, 1.9931303585052491, 1.992877590637207, 1.9936017736434937, 1.9915303264999389, 1.9928937450408934, 1.9911755990600586, 1.9931563230133056, 1.992949102935791, 1.9930351052856445, 1.9907214881134032, 1.9905328282165526, 1.9933595889282227, 1.9914652145385743, 1.99171005859375, 1.9924902798080444, 1.992575800704956, 1.99016278049469, 1.9921005247497559, 1.9918115587615968, 1.9918856573867798, 1.988608092727661, 1.9906916750717163 ], "train_acc": [ 0.23438, 0.24682, 0.24856, 0.24744, 0.2493, 0.25268, 0.25096, 0.25544, 0.25558, 0.25914, 0.2613, 0.2586, 0.26004, 0.26336, 0.2627, 0.26136, 0.26178, 0.26482, 0.26748, 0.26462, 0.26374, 0.26504, 0.26682, 0.26656, 0.26808, 0.26636, 0.26534, 0.26986, 0.26678, 0.26804, 0.27068, 0.27112, 0.2728, 0.27114, 0.27098, 0.27242, 0.27134, 0.269, 0.27124, 0.27356, 0.27462, 0.27154, 0.27048, 0.27332, 0.27462, 0.2742, 0.2748, 0.27136, 0.27366, 0.27372, 0.2756, 0.2745, 0.27688, 0.2753, 0.2738, 0.27568, 0.27534, 0.2752, 0.2746, 0.27426, 0.2754, 0.27652, 0.27636, 0.27514, 0.27602, 0.2772, 0.27466, 0.27792, 0.27762, 0.27772, 0.2748, 0.27744, 0.27738, 0.2778, 0.27648, 0.27882, 0.27752, 0.2773, 0.27792, 0.27792, 0.27564, 0.2783, 0.2765, 0.27966, 0.27638, 0.27776, 0.27614, 0.27868, 0.27806, 0.27834, 0.27876, 0.2785, 0.27914, 0.27822, 0.27976, 0.2771, 0.2782, 0.27906, 0.27974, 0.278 ], "test_acc": [ 0.267, 0.2833, 0.2633, 0.2623, 0.2722, 0.2805, 0.2629, 0.2946, 0.2836, 0.2778, 0.2841, 0.2871, 0.2812, 0.2675, 0.2747, 0.2994, 0.2768, 0.2969, 0.2894, 0.3029, 0.2895, 0.296, 0.2693, 0.2842, 0.2805, 0.2916, 0.2823, 0.2769, 0.2952, 0.2914, 0.2952, 0.2872, 0.2933, 0.2701, 0.2877, 0.2897, 0.2824, 0.2808, 0.2978, 0.2775, 0.2873, 0.2938, 0.2942, 0.2895, 0.2924, 0.2896, 0.2957, 0.2915, 0.294, 0.2914, 0.2868, 0.2964, 0.2861, 0.2952, 0.2823, 0.3038, 0.2965, 0.3008, 0.2928, 0.2991, 0.2963, 0.295, 0.2983, 0.2957, 0.2997, 0.2953, 0.2891, 0.2982, 0.3011, 0.2934, 0.2948, 0.305, 0.2971, 0.3022, 0.2961, 0.2952, 0.3052, 0.2981, 0.2997, 0.2923, 0.2967, 0.299, 0.2967, 0.2969, 0.2986, 0.2949, 0.2994, 0.2954, 0.2994, 0.2989, 0.2964, 0.2977, 0.2943, 0.2971, 0.296, 0.2965, 0.296, 0.2968, 0.2968, 0.2967 ] }, "diagnostics": { "bp_cosine": [ 0.39808189868927, -0.0005002215621061623, -0.00010176157957175747, -0.00011934398207813501, -0.000299518636893481, 0.0004310115473344922, -0.0001575749192852527, -0.00013478680921252817, -0.00013778718130197376, -0.0002350577706238255, -7.549257134087384e-06, -3.051616658922285e-05 ], "perturbation_rho": [ 0.026208851486444473, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "nudging": { "0.001": [ -3.781169652938843e-07, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "0.003": [ -9.867362678050995e-07, 0.0, -1.1175870895385742e-08, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -1.862645149230957e-09, 0.0 ], "0.01": [ -3.6079436540603638e-06, 0.0, -1.30385160446167e-08, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -1.862645149230957e-09, 0.0 ] } }, "drift": { "embed.weight": 330.294227499211, "embed.bias": 259.4924618202614, "blocks.0.ln.weight": 10.952609998429608, "blocks.0.w1.weight": 280.8978034197629, "blocks.0.w1.bias": 260.7432178686956, "blocks.0.w2.weight": 459.75981344197356, "blocks.1.ln.weight": 8.338033915232403, "blocks.1.w1.weight": 303.52507755893413, "blocks.1.w1.bias": 305.583927901598, "blocks.1.w2.weight": 333.8270370300826, "blocks.2.ln.weight": 10.70947460327803, "blocks.2.w1.weight": 423.89849941751544, "blocks.2.w1.bias": 420.3062621668287, "blocks.2.w2.weight": 429.4082983364606, "blocks.3.ln.weight": 8.290240736902255, "blocks.3.w1.weight": 317.09537244180814, "blocks.3.w1.bias": 287.7440748790634, "blocks.3.w2.weight": 298.4352829162209, "blocks.4.ln.weight": 9.625651853706938, "blocks.4.w1.weight": 380.24334326208, "blocks.4.w1.bias": 350.1027945318976, "blocks.4.w2.weight": 351.4361032986557, "blocks.5.ln.weight": 9.64751089559273, "blocks.5.w1.weight": 381.41558228811317, "blocks.5.w1.bias": 348.41973459641565, "blocks.5.w2.weight": 358.39732437217214, "blocks.6.ln.weight": 9.747382743882953, "blocks.6.w1.weight": 390.26899720474756, "blocks.6.w1.bias": 371.3396255333559, "blocks.6.w2.weight": 383.2513961535749, "blocks.7.ln.weight": 8.825122691585458, "blocks.7.w1.weight": 346.68999846263586, "blocks.7.w1.bias": 322.5261718218226, "blocks.7.w2.weight": 338.7459123423596, "blocks.8.ln.weight": 8.098539333504899, "blocks.8.w1.weight": 318.2267389133952, "blocks.8.w1.bias": 297.59966764417567, "blocks.8.w2.weight": 295.4097011432017, "blocks.9.ln.weight": 7.141109927159286, "blocks.9.w1.weight": 262.58513708371515, "blocks.9.w1.bias": 251.1580247152912, "blocks.9.w2.weight": 243.52122031424597, "blocks.10.ln.weight": 10.631727490611157, "blocks.10.w1.weight": 421.8663669312304, "blocks.10.w1.bias": 393.793856832655, "blocks.10.w2.weight": 426.1507190170269, "blocks.11.ln.weight": 11.204882830494729, "blocks.11.w1.weight": 446.3738461267888, "blocks.11.w1.bias": 412.5154467592355, "blocks.11.w2.weight": 426.65837546735577, "out_ln.weight": 0.6667026319780115, "out_head.weight": 9.48630928590299, "out_head.bias": 0.44493816450857304 } }, "state_bridge": { "log": { "train_loss": [ 2.3018941720581054, 2.2664404193878176, 2.241582181625366, 2.2375785102844237, 2.134529279785156, 2.119604638824463, 2.112871742706299, 2.1424258409881594, 2.158468070068359, 2.143250986175537, 2.1479166007995607, 2.141409687652588, 2.1494772535705566, 2.1696451356506348, 2.2107076875305176, 2.216366109008789, 2.2074287937927246, 2.1900475843811034, 2.1331632147216797, 2.130491197357178, 2.1317693274688723, 2.1269900135040283, 2.1228091914367675, 2.099304881210327, 2.085384643859863, 2.085484384918213, 2.087498538970947, 2.0871505451202395, 2.0823512071990966, 2.0860876009368896, 2.086019044647217, 2.0951096993255613, 2.0864567431259156, 2.078469701538086, 2.072162138824463, 2.0764309199523927, 2.0748440283203125, 2.0789254627990723, 2.0826576986694336, 2.0871692906188963, 2.086170325393677, 2.084945154571533, 2.0844455377197266, 2.085197698135376, 2.0900804764556886, 2.082403769226074, 2.082503260040283, 2.0882599214935302, 2.0945081477355956, 2.0928621436309816, 2.091693960571289, 2.101140612335205, 2.095073269882202, 2.0920650315856935, 2.091407716522217, 2.092963116760254, 2.102848686065674, 2.10209814743042, 2.100728988342285, 2.1085473413085936, 2.1115043045043946, 2.1238191069030763, 2.1308215964508057, 2.1344855577087403, 2.1272054124450683, 2.1218687773132325, 2.12194469039917, 2.127710237426758, 2.1268114820861816, 2.1254331665039063, 2.1245197431182863, 2.125965220489502, 2.12684192817688, 2.1273793972015382, 2.128570475311279, 2.1321925525665284, 2.13282696762085, 2.1370943976593018, 2.1397155253601072, 2.143300206756592, 2.1476288512420654, 2.152116884613037, 2.1553422878265383, 2.154549729385376, 2.155469595031738, 2.155468685760498, 2.15419854057312, 2.1527544927978517, 2.151362074584961, 2.1510060634613035, 2.1480546588134763, 2.1468420223999023, 2.1457022499084473, 2.1455462963104246, 2.144765656814575, 2.146289216003418, 2.142656379470825, 2.1456896996307373, 2.1448883264923095, 2.144138221282959 ], "train_acc": [ 0.13994, 0.14018, 0.1448, 0.1448, 0.16884, 0.17316, 0.1748, 0.1631, 0.15902, 0.1599, 0.15558, 0.16126, 0.16064, 0.15736, 0.1483, 0.14638, 0.15088, 0.15526, 0.16726, 0.16864, 0.1672, 0.17562, 0.1804, 0.1857, 0.19408, 0.19066, 0.1875, 0.19146, 0.19404, 0.18652, 0.1868, 0.18316, 0.1862, 0.19086, 0.19266, 0.18878, 0.19202, 0.18684, 0.18708, 0.18076, 0.18768, 0.18408, 0.18372, 0.18428, 0.18508, 0.18442, 0.1863, 0.18434, 0.18142, 0.17946, 0.18168, 0.18058, 0.1805, 0.18198, 0.18268, 0.1812, 0.17942, 0.17946, 0.18054, 0.17744, 0.17834, 0.17516, 0.17416, 0.17206, 0.17546, 0.17682, 0.17458, 0.17444, 0.1757, 0.17654, 0.1778, 0.17268, 0.1757, 0.1751, 0.17666, 0.1775, 0.17746, 0.17622, 0.17406, 0.17472, 0.17496, 0.17596, 0.17324, 0.17394, 0.17302, 0.1713, 0.17072, 0.17288, 0.1718, 0.17242, 0.1725, 0.17408, 0.1736, 0.17198, 0.17344, 0.17254, 0.17392, 0.17376, 0.17466, 0.1739 ], "test_acc": [ 0.0608, 0.1367, 0.1165, 0.1638, 0.1577, 0.176, 0.1529, 0.136, 0.1352, 0.1245, 0.1322, 0.137, 0.1201, 0.1178, 0.1187, 0.1118, 0.1187, 0.1373, 0.1421, 0.1151, 0.1412, 0.1284, 0.1374, 0.1796, 0.1927, 0.1818, 0.1859, 0.1762, 0.1865, 0.1827, 0.1708, 0.1826, 0.1859, 0.191, 0.1833, 0.1858, 0.1875, 0.1902, 0.1859, 0.1872, 0.1894, 0.193, 0.1882, 0.1783, 0.1792, 0.179, 0.1836, 0.1618, 0.1749, 0.1881, 0.1879, 0.1757, 0.1705, 0.1868, 0.1871, 0.1811, 0.1727, 0.16, 0.1799, 0.1524, 0.1648, 0.1445, 0.1739, 0.1721, 0.1828, 0.1857, 0.1719, 0.1722, 0.1716, 0.1694, 0.1512, 0.1719, 0.1704, 0.1747, 0.1718, 0.1712, 0.173, 0.175, 0.1733, 0.1749, 0.165, 0.1672, 0.1708, 0.161, 0.155, 0.1605, 0.1629, 0.1598, 0.1625, 0.1628, 0.1625, 0.1609, 0.1603, 0.1616, 0.1607, 0.1644, 0.1622, 0.1621, 0.1639, 0.1637 ], "state_pred_error": [ 0.4747012776184082, 0.8295380383682252, 0.21469590087890625, 0.12448142728805542, 0.08570515047073364, 0.08470476587295532, 0.12634123967647554, 0.09733817555308343, 0.0353814222663641, 0.03443111445069313, 0.03137257686555386, 0.029351851464509963, 0.03382749609231949, 0.03600793175697327, 0.028298149601221085, 0.02450420175075531, 0.025809087952375413, 0.033925389347076416, 0.039575903968811034, 0.03546395682752133, 0.03896381706476212, 0.035592052783966065, 0.020547916915416718, 0.027013523270487786, 0.024955789625644684, 0.01776892081618309, 0.010803042232394219, 0.008741003857254981, 0.009176803540587425, 0.00826898938536644, 0.007538356708586216, 0.0065028233194351194, 0.006696901108622551, 0.006364495917409658, 0.004348094679266215, 0.0027382008469104767, 0.001712799895554781, 0.001062507140468806, 0.0006733776653371752, 0.0004975279747322202, 0.0005460627448558807, 0.0007091301748156548, 0.0008156654949113727, 0.0005741439528390766, 0.0004632441856153309, 0.0004077366246469319, 0.00037149273524992167, 0.00032670746902935206, 0.0003252365568652749, 0.0002759585575200617, 0.00019474619667977095, 0.00015623675198294222, 0.00016898068460635842, 0.00014236534488853067, 0.00010944471210241317, 8.656780364457518e-05, 8.127526033436879e-05, 7.092417092993855e-05, 7.672614848706871e-05, 6.656669022282586e-05, 6.597209698753432e-05, 6.993545684730634e-05, 7.107698564883322e-05, 7.049443874042482e-05, 6.95053919684142e-05, 6.41730681899935e-05, 6.155454550869764e-05, 6.015078729018569e-05, 5.4422928835265336e-05, 4.830317224375904e-05, 4.409612072864547e-05, 3.9640857164049524e-05, 3.758983974112198e-05, 3.4887928910320625e-05, 3.156908231205307e-05, 2.8848007169435732e-05, 2.6238742939895017e-05, 2.2889468831708655e-05, 2.021120209712535e-05, 1.8343317908002065e-05, 1.5500704940350262e-05, 1.386145944125019e-05, 1.3288812919927295e-05, 1.3635638843989e-05, 1.4378732023760676e-05, 1.5836126625945327e-05, 1.6757641757139935e-05, 1.6754487288417295e-05, 1.670292949071154e-05, 1.5964577784761785e-05, 1.5669193891226315e-05, 1.4928079603705555e-05, 1.4118930798722432e-05, 1.3290679465862922e-05, 1.259699832968181e-05, 1.169607178599108e-05, 1.0905416535097175e-05, 9.99030330684036e-06, 9.378302599070595e-06, 8.833123445947422e-06 ] }, "diagnostics": { "bp_cosine": [ 0.07415470480918884, 0.005267417058348656, 0.006899588741362095, 0.009260526858270168, 0.014165668748319149, 0.007070831023156643, 0.003572859801352024, 0.0013162376126274467, 0.01693638414144516, 0.011615054681897163, 0.01297210343182087, 0.019879184663295746 ], "perturbation_rho": [ 0.06713278591632843, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "nudging": { "0.001": [ -3.7383288145065308e-06, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "0.003": [ -1.1199153959751129e-05, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "0.01": [ -3.7682242691516876e-05, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ] } }, "drift": { "embed.weight": 16.995931312362433, "embed.bias": 107.50455214257622, "blocks.0.ln.weight": 1.1747809607171977, "blocks.0.w1.weight": 87.48144222271058, "blocks.0.w1.bias": 108.71150020387606, "blocks.0.w2.weight": 231.25407676027953, "blocks.1.ln.weight": 2.5949076659471078, "blocks.1.w1.weight": 171.7945773092806, "blocks.1.w1.bias": 169.71996864021074, "blocks.1.w2.weight": 213.0622571876725, "blocks.2.ln.weight": 3.47458321612997, "blocks.2.w1.weight": 89.72539771461543, "blocks.2.w1.bias": 64.43223038220326, "blocks.2.w2.weight": 130.15474280855415, "blocks.3.ln.weight": 3.2490652039574544, "blocks.3.w1.weight": 125.1742859456183, "blocks.3.w1.bias": 103.82426516529186, "blocks.3.w2.weight": 122.83893982946046, "blocks.4.ln.weight": 3.143539271549031, "blocks.4.w1.weight": 124.0384309768722, "blocks.4.w1.bias": 100.70299850760337, "blocks.4.w2.weight": 149.19128092745697, "blocks.5.ln.weight": 3.8369736269462678, "blocks.5.w1.weight": 70.0473609489046, "blocks.5.w1.bias": 55.88760753160443, "blocks.5.w2.weight": 86.95978496337405, "blocks.6.ln.weight": 2.788527931096308, "blocks.6.w1.weight": 68.53157012655858, "blocks.6.w1.bias": 60.09690095738387, "blocks.6.w2.weight": 89.81505023786389, "blocks.7.ln.weight": 6.399420463964318, "blocks.7.w1.weight": 211.5238472865523, "blocks.7.w1.bias": 193.8907626146654, "blocks.7.w2.weight": 198.88821232136138, "blocks.8.ln.weight": 3.718593242696815, "blocks.8.w1.weight": 131.87660631337303, "blocks.8.w1.bias": 117.13821442329933, "blocks.8.w2.weight": 175.03912755183418, "blocks.9.ln.weight": 2.399620577204472, "blocks.9.w1.weight": 62.45119981781294, "blocks.9.w1.bias": 48.049978443761816, "blocks.9.w2.weight": 88.05873456328695, "blocks.10.ln.weight": 4.012307554325392, "blocks.10.w1.weight": 150.00438163426782, "blocks.10.w1.bias": 134.3950208692614, "blocks.10.w2.weight": 137.82521611752918, "blocks.11.ln.weight": 2.857283266360585, "blocks.11.w1.weight": 108.3646369610001, "blocks.11.w1.bias": 100.36750299157434, "blocks.11.w2.weight": 117.58094348226459, "out_ln.weight": 0.45424689776008886, "out_head.weight": 5.496813200883646, "out_head.bias": 3.0092640143628198 } }, "credit_bridge": { "log": { "train_loss": [ 2.080038804626465, 2.0486898488616943, 2.046282894744873, 2.046656800842285, 2.039552648086548, 2.0396751835632325, 2.037381534881592, 2.035384338226318, 2.035809118041992, 2.0331781942749023, 2.0284877285385132, 2.027495804901123, 2.026125898208618, 2.020480899810791, 2.0207955992126463, 2.0185290005874634, 2.017150591125488, 2.0164879455566407, 2.0157563932037355, 2.0127871045684813, 2.0112040578460695, 2.010815860519409, 2.009171763458252, 2.007995718460083, 2.0075271183013914, 2.006510923538208, 2.003063318405151, 2.004764850692749, 2.0030145164489745, 2.001830623397827, 2.0015824402618407, 2.0029561541748047, 2.0005920767593386, 1.9996838995361328, 2.0028357711029052, 2.0006233686065675, 2.002561650390625, 2.0042890097808836, 2.0027454135513305, 2.0027183213806152, 2.0006818413543703, 2.0015592734146117, 1.9975921591949464, 1.9961159555053711, 1.9909970508575439, 1.9920473384857178, 1.9899114678192138, 1.9886371549987794, 1.9901284057998658, 1.9869096327972413, 1.9846498122406007, 1.986077251586914, 1.9829593098449707, 1.9857685099792481, 1.9820374918365478, 1.9831326490783692, 1.9809515007781981, 1.9830830813980103, 1.9793231730651855, 1.9804900155639649, 1.9807616325378419, 1.9796239783096314, 1.9790023094177247, 1.976966079940796, 1.978210344467163, 1.9777095768737794, 1.9761795248794556, 1.9757038269042968, 1.9747423070907593, 1.9738406778717041, 1.9735038549041748, 1.97461932346344, 1.9734378964233399, 1.9701263136291505, 1.9708808054351807, 1.974146450958252, 1.9722303295135497, 1.9706154293823241, 1.968834009361267, 1.970496709022522, 1.9716806288909912, 1.9713043495941163, 1.9710750988769532, 1.9707429261016847, 1.9708539364242554, 1.9677796157455445, 1.969852265663147, 1.9710520673370362, 1.970287728881836, 1.9704584929656983, 1.9702265856552124, 1.9696108251571656, 1.9689367693328856, 1.9707266571044921, 1.967924552001953, 1.9694965203857422, 1.9685778493881225, 1.9698442002868652, 1.9704585931015015, 1.9684682028961182 ], "train_acc": [ 0.23092, 0.24236, 0.24134, 0.2427, 0.24468, 0.24838, 0.25104, 0.24966, 0.24992, 0.2533, 0.25824, 0.25604, 0.25956, 0.25866, 0.25884, 0.26064, 0.26036, 0.26042, 0.26344, 0.26316, 0.26514, 0.26496, 0.26788, 0.26702, 0.26586, 0.2673, 0.27052, 0.26806, 0.26714, 0.26944, 0.26906, 0.26936, 0.27068, 0.27078, 0.26864, 0.26786, 0.26666, 0.26666, 0.26544, 0.26578, 0.26854, 0.26384, 0.26708, 0.26708, 0.2699, 0.26658, 0.2711, 0.26876, 0.26902, 0.27082, 0.27148, 0.266, 0.27216, 0.26798, 0.27106, 0.26994, 0.27048, 0.26962, 0.2742, 0.27182, 0.27056, 0.27532, 0.27016, 0.2717, 0.2714, 0.27288, 0.274, 0.27434, 0.27378, 0.27492, 0.27316, 0.27746, 0.27542, 0.27628, 0.27614, 0.27594, 0.27926, 0.27974, 0.28028, 0.27698, 0.27918, 0.28146, 0.27674, 0.27706, 0.2764, 0.27804, 0.27652, 0.27794, 0.27888, 0.27896, 0.27918, 0.2798, 0.27814, 0.27842, 0.2812, 0.28084, 0.27868, 0.2795, 0.27768, 0.28226 ], "test_acc": [ 0.258, 0.2756, 0.2609, 0.2627, 0.2583, 0.2669, 0.2661, 0.2481, 0.2517, 0.278, 0.2729, 0.2822, 0.2748, 0.2742, 0.2787, 0.2778, 0.2707, 0.2794, 0.2583, 0.2854, 0.2723, 0.2857, 0.2747, 0.2855, 0.2884, 0.2972, 0.3, 0.2951, 0.278, 0.3009, 0.2933, 0.2909, 0.2962, 0.2866, 0.2697, 0.2857, 0.2673, 0.2944, 0.2914, 0.2886, 0.2744, 0.2794, 0.286, 0.2811, 0.2905, 0.2843, 0.2743, 0.2756, 0.279, 0.2945, 0.2845, 0.2711, 0.2852, 0.2824, 0.2842, 0.2934, 0.2924, 0.287, 0.2801, 0.2987, 0.282, 0.2911, 0.2793, 0.2753, 0.2963, 0.2823, 0.2872, 0.2971, 0.2958, 0.2753, 0.285, 0.2899, 0.2999, 0.2834, 0.2931, 0.2934, 0.2869, 0.2886, 0.2916, 0.2905, 0.2822, 0.2883, 0.2927, 0.3001, 0.295, 0.2954, 0.2931, 0.2848, 0.2906, 0.2907, 0.291, 0.2922, 0.2929, 0.2921, 0.2908, 0.2913, 0.29, 0.2895, 0.2894, 0.2893 ], "value_loss": [ 0.4853640641069412, 0.13146528634786606, 0.10278774485111236, 0.10451403197526932, 0.09186984123706818, 0.09725872883081436, 0.08755186069607734, 0.07744781212091446, 0.08400812898874283, 0.0792426191854477, 0.07378534253954887, 0.06448388241052627, 0.06442826303482056, 0.062765171585083, 0.05268099952340126, 0.046547150465250015, 0.04977830931901932, 0.046939543528556824, 0.04594481763601303, 0.04108059468030929, 0.04083769999027252, 0.04402205483198166, 0.04000264639854431, 0.03796569113850594, 0.035944590538740155, 0.03808709916114807, 0.03455143286049366, 0.038163829424381256, 0.03331702991783619, 0.033281468888521196, 0.02748794436097145, 0.030440516521930694, 0.028207264231443405, 0.02704140617221594, 0.026097478928565978, 0.025327355717718603, 0.024099903733730315, 0.027659756012558937, 0.022987231121063233, 0.020967545130252837, 0.021579559311568738, 0.020051995359063147, 0.01838304534137249, 0.02002642317146063, 0.017698023419380187, 0.018389571142196654, 0.013522476999759675, 0.01383736302614212, 0.014502173573970795, 0.016614098087251186, 0.013939322325885297, 0.014496620377600192, 0.010598833594471216, 0.012348202850222588, 0.011290375513732433, 0.009235630955398083, 0.011637351908683776, 0.00849102252393961, 0.009930187560021878, 0.0093223687428236, 0.008464423409998418, 0.007654866740107537, 0.007954280150234699, 0.00844777645289898, 0.008321732766479253, 0.007419437230825424, 0.008567752386033535, 0.007364120395183563, 0.006373357061594725, 0.00595162553101778, 0.006675690564513207, 0.0059091412272304295, 0.0058950428107380866, 0.005543788010179997, 0.004646000351905823, 0.004232099659517407, 0.00391991232007742, 0.003104630061686039, 0.0029366533225774765, 0.0028203235547989607, 0.0027620538933947683, 0.002386382454186678, 0.0019722532033920287, 0.001968352645486593, 0.001672238507643342, 0.0015232293362915515, 0.0017160395857319236, 0.0012631227846443653, 0.0011414727784320711, 0.0009458377653546632, 0.0011035545341856778, 0.0009393164394423365, 0.001123441938534379, 0.0008937438951432705, 0.0009866366090252995, 0.0011745557901449502, 0.0006501818283647299, 0.0009478944011032582, 0.001031599847562611, 0.0008783576951175928 ] }, "diagnostics": { "bp_cosine": [ 0.4526920020580292, 0.006497844588011503, 0.01631792262196541, 0.01694033294916153, 0.017034951597452164, 0.01775282248854637, 0.01775078848004341, 0.017723549157381058, 0.01778416708111763, 0.01786961778998375, 0.017997052520513535, 0.017978299409151077 ], "perturbation_rho": [ -0.054250095039606094, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "nudging": { "0.001": [ -8.353963494300842e-07, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "0.003": [ -2.439599484205246e-06, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "0.01": [ -7.801689207553864e-06, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ] } }, "drift": { "embed.weight": 193.53208266596525, "embed.bias": 115.78380404718655, "blocks.0.ln.weight": 8.867635038294205, "blocks.0.w1.weight": 196.51094816310516, "blocks.0.w1.bias": 74.82582896315267, "blocks.0.w2.weight": 317.40620630343955, "blocks.1.ln.weight": 5.764116391727315, "blocks.1.w1.weight": 210.05221683875882, "blocks.1.w1.bias": 200.7126628347933, "blocks.1.w2.weight": 235.05000374612118, "blocks.2.ln.weight": 7.735844453849748, "blocks.2.w1.weight": 328.46033970353545, "blocks.2.w1.bias": 322.5548152476072, "blocks.2.w2.weight": 334.35863169713303, "blocks.3.ln.weight": 5.5455094513045, "blocks.3.w1.weight": 213.2636693536328, "blocks.3.w1.bias": 189.3065053401033, "blocks.3.w2.weight": 201.9218429504037, "blocks.4.ln.weight": 6.963102540839665, "blocks.4.w1.weight": 281.1829353200361, "blocks.4.w1.bias": 254.14827278048986, "blocks.4.w2.weight": 252.72823154554072, "blocks.5.ln.weight": 6.651053338270696, "blocks.5.w1.weight": 265.808423173078, "blocks.5.w1.bias": 237.94000901426764, "blocks.5.w2.weight": 246.35312384551574, "blocks.6.ln.weight": 6.796230689261986, "blocks.6.w1.weight": 277.37913499866477, "blocks.6.w1.bias": 258.5087504829616, "blocks.6.w2.weight": 266.3321474533456, "blocks.7.ln.weight": 6.190933150629552, "blocks.7.w1.weight": 247.84752961611798, "blocks.7.w1.bias": 225.69508996072392, "blocks.7.w2.weight": 237.71747748473527, "blocks.8.ln.weight": 5.572316870464946, "blocks.8.w1.weight": 220.49723448227533, "blocks.8.w1.bias": 201.84500185584645, "blocks.8.w2.weight": 203.8926634145943, "blocks.9.ln.weight": 4.956248620217523, "blocks.9.w1.weight": 176.87866597616605, "blocks.9.w1.bias": 165.25106759218716, "blocks.9.w2.weight": 166.2448646005026, "blocks.10.ln.weight": 7.452869833755125, "blocks.10.w1.weight": 298.53765102170973, "blocks.10.w1.bias": 273.411202411899, "blocks.10.w2.weight": 295.16401679878913, "blocks.11.ln.weight": 7.774196063602293, "blocks.11.w1.weight": 313.80618162900083, "blocks.11.w1.bias": 283.94177751762567, "blocks.11.w2.weight": 294.8394911964541, "out_ln.weight": 0.5766241941627794, "out_head.weight": 8.360089920565844, "out_head.bias": 1.5207819852306046 } } }, "config": { "dataset": "cifar10", "d_hidden": 512, "num_blocks": 12, "batch_size": 128, "epochs": 100, "lr": 0.001, "lr_fb": 0.001, "wd": 0.01, "lam": 0.1, "K": 4, "sigma_bridge": 0.05, "ema_momentum": 0.995, "term_grad_weight": 1.0, "seeds": [ 456 ], "gpu": 0, "output_dir": "results/cifar10_seed456", "num_classes": 10 } }