amirali1985 commited on
Commit
1e6658c
·
verified ·
1 Parent(s): eec6d9b

Upload add_sub_sorl_v1_abs10_K1_25K

Browse files
add_sub_sorl_v1_abs10_K1_25K/metrics.json CHANGED
@@ -70,444 +70,936 @@
70
  3719,
71
  3769,
72
  3819,
73
- 3869
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
74
  ],
75
  "loss": [
76
- 9.190461158752441,
77
- 5.166501045227051,
78
- 3.2305479049682617,
79
- 3.0906858444213867,
80
- 2.994469165802002,
81
- 3.048895835876465,
82
- 3.2591943740844727,
83
- 1.6000936031341553,
84
- -0.03324759006500244,
85
- -3.219552516937256,
86
- -4.197554588317871,
87
- -2.6743619441986084,
88
- -2.2282357215881348,
89
- -1.7437527179718018,
90
- -1.24261474609375,
91
- -1.1773107051849365,
92
- -0.9963365793228149,
93
- -0.18717944622039795,
94
- -0.994360089302063,
95
- -0.17535710334777832,
96
- -0.5866737961769104,
97
- -0.32630622386932373,
98
- -0.1583651304244995,
99
- -0.5506547689437866,
100
- -0.6750422716140747,
101
- -0.5237396955490112,
102
- -0.5621165037155151,
103
- -0.07628041505813599,
104
- -0.1960809826850891,
105
- -0.5238226652145386,
106
- -0.45479297637939453,
107
- -0.5785545706748962,
108
- -0.572192907333374,
109
- -0.7545036673545837,
110
- -0.39050161838531494,
111
- -0.3878746032714844,
112
- -0.246707946062088,
113
- -0.3712904155254364,
114
- -0.7410520911216736,
115
- -0.7477802634239197,
116
- -0.9374204277992249,
117
- -0.6002537608146667,
118
- -0.5804862380027771,
119
- -0.7307106256484985,
120
- -1.1449896097183228,
121
- -0.7023072242736816,
122
- -0.2751615643501282,
123
- -0.8681702613830566,
124
- -0.6093825101852417,
125
- -0.3929210901260376,
126
- -0.7827325463294983,
127
- -0.7987732887268066,
128
- -0.6832472085952759,
129
- -0.5688872337341309,
130
- -0.6018052101135254,
131
- -0.8418792486190796,
132
- -0.7785307765007019,
133
- -0.337721049785614,
134
- -0.5132121443748474,
135
- -0.4073343873023987,
136
- -0.3655291199684143,
137
- -0.6873089075088501,
138
- -0.4807700514793396,
139
- -0.39133989810943604,
140
- -0.4098546802997589,
141
- -0.797407865524292,
142
- -0.44859200716018677,
143
- -0.4712650775909424,
144
- -0.3992188572883606,
145
- -0.4064672291278839
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
146
  ],
147
  "base_loss": [
148
- 7.707293510437012,
149
- 4.35805606842041,
150
- 2.015561819076538,
151
- 1.9516109228134155,
152
- 1.93378746509552,
153
- 1.8740915060043335,
154
- 1.8301212787628174,
155
- 1.8807543516159058,
156
- 1.8543447256088257,
157
- 1.7816431522369385,
158
- 1.449736475944519,
159
- 1.0750778913497925,
160
- 0.8969014286994934,
161
- 0.7201328873634338,
162
- 0.5857338309288025,
163
- 0.515350341796875,
164
- 0.4651491045951843,
165
- 0.402709424495697,
166
- 0.44452181458473206,
167
- 0.35987958312034607,
168
- 0.3687458038330078,
169
- 0.309616357088089,
170
- 0.28466305136680603,
171
- 0.2800248861312866,
172
- 0.37805360555648804,
173
- 0.349456250667572,
174
- 0.2721202075481415,
175
- 0.1941368132829666,
176
- 0.2818446159362793,
177
- 0.24511845409870148,
178
- 0.22365573048591614,
179
- 0.2581332325935364,
180
- 0.22468924522399902,
181
- 0.2866680920124054,
182
- 0.19988836348056793,
183
- 0.21447350084781647,
184
- 0.19538240134716034,
185
- 0.19758661091327667,
186
- 0.17485487461090088,
187
- 0.1870802789926529,
188
- 0.1876666396856308,
189
- 0.1586993783712387,
190
- 0.15935318171977997,
191
- 0.1772172749042511,
192
- 0.20511488616466522,
193
- 0.14057905972003937,
194
- 0.14634256064891815,
195
- 0.17220208048820496,
196
- 0.15756499767303467,
197
- 0.12289166450500488,
198
- 0.13189402222633362,
199
- 0.12689675390720367,
200
- 0.1268821805715561,
201
- 0.0976928099989891,
202
- 0.09254758805036545,
203
- 0.1167963370680809,
204
- 0.12230370193719864,
205
- 0.061435870826244354,
206
- 0.10646288096904755,
207
- 0.06817925721406937,
208
- 0.07916043698787689,
209
- 0.09582412242889404,
210
- 0.07675866037607193,
211
- 0.06337268650531769,
212
- 0.06639906018972397,
213
- 0.10464208573102951,
214
- 0.06369388103485107,
215
- 0.07750404626131058,
216
- 0.05772814899682999,
217
- 0.0618344210088253
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
218
  ],
219
  "info_loss": [
220
- -0.5703163146972656,
221
- -0.166015625,
222
- -0.06974220275878906,
223
- -0.0745776891708374,
224
- -0.08152985572814941,
225
- -0.06991314888000488,
226
- -0.04411435127258301,
227
- -0.21549618244171143,
228
- -0.37584030628204346,
229
- -0.6874351501464844,
230
- -0.7523373961448669,
231
- -0.557295024394989,
232
- -0.4853973090648651,
233
- -0.40835925936698914,
234
- -0.3295533061027527,
235
- -0.3078162670135498,
236
- -0.2742995619773865,
237
- -0.18321914970874786,
238
- -0.2575337886810303,
239
- -0.15976710617542267,
240
- -0.1911141574382782,
241
- -0.15414360165596008,
242
- -0.13090449571609497,
243
- -0.15989911556243896,
244
- -0.18713147938251495,
245
- -0.1621924787759781,
246
- -0.15000823140144348,
247
- -0.09506124258041382,
248
- -0.10266301035881042,
249
- -0.13723227381706238,
250
- -0.12602582573890686,
251
- -0.13937442004680634,
252
- -0.1260860562324524,
253
- -0.14851446449756622,
254
- -0.10906409472227097,
255
- -0.09848323464393616,
256
- -0.07746206969022751,
257
- -0.0828794315457344,
258
- -0.11464187502861023,
259
- -0.11303246766328812,
260
- -0.1308683156967163,
261
- -0.09133747220039368,
262
- -0.09089986979961395,
263
- -0.1075134426355362,
264
- -0.1480427086353302,
265
- -0.09612017124891281,
266
- -0.054484449326992035,
267
- -0.1174512505531311,
268
- -0.08855501562356949,
269
- -0.06146596744656563,
270
- -0.10301853716373444,
271
- -0.10187290608882904,
272
- -0.09169016033411026,
273
- -0.07557268440723419,
274
- -0.0797661542892456,
275
- -0.10500068217515945,
276
- -0.0984283983707428,
277
- -0.047317616641521454,
278
- -0.06959307193756104,
279
- -0.05479596555233002,
280
- -0.05312419682741165,
281
- -0.08555228263139725,
282
- -0.06329430639743805,
283
- -0.052405696362257004,
284
- -0.05460964888334274,
285
- -0.09755084663629532,
286
- -0.05779761075973511,
287
- -0.06151659041643143,
288
- -0.05198967456817627,
289
- -0.05291672796010971
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
290
  ],
291
  "abs_loss": [
292
- 2.2391064167022705,
293
- 1.903133749961853,
294
- 1.8471856117248535,
295
- 1.870505690574646,
296
- 1.8405386209487915,
297
- 1.8321367502212524,
298
- 1.8374236822128296,
299
- 1.8635932207107544,
300
- 1.8500641584396362,
301
- 1.8130455017089844,
302
- 1.789480209350586,
303
- 1.4839378595352173,
304
- 1.1312556266784668,
305
- 0.8757233619689941,
306
- 0.6512569785118103,
307
- 0.6017574667930603,
308
- 0.5022904872894287,
309
- 0.46593570709228516,
310
- 0.40201544761657715,
311
- 0.35612478852272034,
312
- 0.29723620414733887,
313
- 0.2846054136753082,
314
- 0.2813468277454376,
315
- 0.2518315017223358,
316
- 0.23474085330963135,
317
- 0.1778661459684372,
318
- 0.22422951459884644,
319
- 0.22165711224079132,
320
- 0.1990116387605667,
321
- 0.21514342725276947,
322
- 0.20219220221042633,
323
- 0.18571658432483673,
324
- 0.1271841675043106,
325
- 0.12184059619903564,
326
- 0.13935674726963043,
327
- 0.1843566745519638,
328
- 0.11972486972808838,
329
- 0.11449414491653442,
330
- 0.1205715611577034,
331
- 0.07506386190652847,
332
- 0.07510096579790115,
333
- 0.07919437438249588,
334
- 0.09589538723230362,
335
- 0.06470391154289246,
336
- 0.06804872304201126,
337
- 0.06406687945127487,
338
- 0.07336432486772537,
339
- 0.04453839734196663,
340
- 0.06374714523553848,
341
- 0.0641837790608406,
342
- 0.051210492849349976,
343
- 0.06126769259572029,
344
- 0.05609944835305214,
345
- 0.045193735510110855,
346
- 0.06044153496623039,
347
- 0.03156470134854317,
348
- 0.04949544370174408,
349
- 0.04925629124045372,
350
- 0.037908535450696945,
351
- 0.052750349044799805,
352
- 0.05024566873908043,
353
- 0.0473237968981266,
354
- 0.056845005601644516,
355
- 0.0363769493997097,
356
- 0.03889637067914009,
357
- 0.04386458173394203,
358
- 0.030996553599834442,
359
- 0.02613372914493084,
360
- 0.025389796122908592,
361
- 0.03995569422841072
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
362
  ],
363
  "zipf_loss": [
364
- 6.962420463562012,
365
- 2.278287887573242,
366
- 1.727689504623413,
367
- 1.6978013515472412,
368
- 1.6919262409210205,
369
- 1.690722107887268,
370
- 1.686474084854126,
371
- 1.6879417896270752,
372
- 1.6858043670654297,
373
- 1.6918517351150513,
374
- 1.6971348524093628,
375
- 1.675116777420044,
376
- 1.6157103776931763,
377
- 1.5321345329284668,
378
- 1.4020588397979736,
379
- 1.3253259658813477,
380
- 1.2312809228897095,
381
- 1.195708990097046,
382
- 1.0962542295455933,
383
- 1.0268218517303467,
384
- 0.9259983897209167,
385
- 0.877052903175354,
386
- 0.8378820419311523,
387
- 0.7431282997131348,
388
- 0.7947447299957275,
389
- 0.7309422492980957,
390
- 0.6434226036071777,
391
- 0.6580294966697693,
392
- 0.5288034081459045,
393
- 0.5818673372268677,
394
- 0.561590313911438,
395
- 0.5384848713874817,
396
- 0.45125994086265564,
397
- 0.4317888617515564,
398
- 0.4863152503967285,
399
- 0.3640486001968384,
400
- 0.3205578625202179,
401
- 0.24846789240837097,
402
- 0.21845467388629913,
403
- 0.18795782327651978,
404
- 0.17608600854873657,
405
- 0.14650212228298187,
406
- 0.15956974029541016,
407
- 0.16073615849018097,
408
- 0.12351780384778976,
409
- 0.11190879344940186,
410
- 0.11600394546985626,
411
- 0.12968632578849792,
412
- 0.11222794651985168,
413
- 0.0924285426735878,
414
- 0.11043774336576462,
415
- 0.08693233132362366,
416
- 0.10116228461265564,
417
- 0.0846274122595787,
418
- 0.0972646176815033,
419
- 0.08817484229803085,
420
- 0.07849989831447601,
421
- 0.06909362971782684,
422
- 0.07246481627225876,
423
- 0.06717102229595184,
424
- 0.08152782917022705,
425
- 0.06765741109848022,
426
- 0.06972979009151459,
427
- 0.06570670753717422,
428
- 0.0659530982375145,
429
- 0.06907197833061218,
430
- 0.06259053200483322,
431
- 0.06378341466188431,
432
- 0.0604107640683651,
433
- 0.05687008425593376
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
434
  ],
435
  "denoise_loss": [],
436
  "ortho_loss": [
437
- 0.4214286804199219,
438
- 0.18882839381694794,
439
- 0.1212303414940834,
440
- 0.11552999913692474,
441
- 0.08706260472536087,
442
- 0.09062568098306656,
443
- 0.08478324860334396,
444
- 0.08018892258405685,
445
- 0.09239780157804489,
446
- 0.1168321818113327,
447
- 0.13350531458854675,
448
- 0.15595880150794983,
449
- 0.1914532333612442,
450
- 0.1955980360507965,
451
- 0.20734401047229767,
452
- 0.2079848051071167,
453
- 0.22458618879318237,
454
- 0.217812642455101,
455
- 0.22359661757946014,
456
- 0.22261333465576172,
457
- 0.21620140969753265,
458
- 0.2151515632867813,
459
- 0.2155912071466446,
460
- 0.21206055581569672,
461
- 0.21063639223575592,
462
- 0.21173329651355743,
463
- 0.2016972005367279,
464
- 0.1977507472038269,
465
- 0.19506840407848358,
466
- 0.19696591794490814,
467
- 0.17341327667236328,
468
- 0.16453030705451965,
469
- 0.18576037883758545,
470
- 0.1828179508447647,
471
- 0.1950673907995224,
472
- 0.1837579756975174,
473
- 0.1741635948419571,
474
- 0.17811855673789978,
475
- 0.16840724647045135,
476
- 0.16616372764110565,
477
- 0.17120879888534546,
478
- 0.16129401326179504,
479
- 0.16783320903778076,
480
- 0.16347329318523407,
481
- 0.16215452551841736,
482
- 0.16703131794929504,
483
- 0.16951780021190643,
484
- 0.16996146738529205,
485
- 0.16912420094013214,
486
- 0.1748487651348114,
487
- 0.1752723902463913,
488
- 0.18083271384239197,
489
- 0.17920422554016113,
490
- 0.18311761319637299,
491
- 0.18380235135555267,
492
- 0.18301932513713837,
493
- 0.1814393401145935,
494
- 0.18424195051193237,
495
- 0.18770591914653778,
496
- 0.18873360753059387,
497
- 0.18713155388832092,
498
- 0.18882162868976593,
499
- 0.18982088565826416,
500
- 0.18827897310256958,
501
- 0.18959622085094452,
502
- 0.18957337737083435,
503
- 0.1906057596206665,
504
- 0.1932428628206253,
505
- 0.19275683164596558,
506
- 0.1928425282239914
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
507
  ],
508
  "lr": [
509
- 3.350427350427351e-05,
510
- 6.76923076923077e-05,
 
 
511
  8e-05,
512
  8e-05,
513
  8e-05,
@@ -548,34 +1040,102 @@
548
  8e-05,
549
  8e-05,
550
  8e-05,
551
- 7.989559715264963e-05,
552
- 7.752280516741365e-05,
553
- 7.51500131821777e-05,
554
- 7.277722119694173e-05,
555
- 7.040442921170577e-05,
556
- 6.803163722646982e-05,
557
- 6.565884524123386e-05,
558
- 6.134036382810441e-05,
559
- 5.896757184286844e-05,
560
- 5.6594779857632485e-05,
561
- 5.422198787239652e-05,
562
- 5.184919588716057e-05,
563
- 4.947640390192461e-05,
564
- 4.710361191668863e-05,
565
- 4.278513050355919e-05,
566
- 4.041233851832322e-05,
567
- 3.803954653308728e-05,
568
- 3.566675454785132e-05,
569
- 3.329396256261535e-05,
570
- 3.092117057737938e-05,
571
- 2.854837859214342e-05,
572
- 2.4229897179013967e-05,
573
- 2.1857105193778026e-05,
574
- 1.9484313208542057e-05,
575
- 1.7111521223306092e-05,
576
- 1.4738729238070129e-05,
577
- 1.2365937252834164e-05,
578
- 9.993145267598198e-06
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
579
  ],
580
  "emb_lr": [],
581
  "eval_step": [
@@ -588,7 +1148,17 @@
588
  2696,
589
  3087,
590
  3478,
591
- 3869
 
 
 
 
 
 
 
 
 
 
592
  ],
593
  "eval_accuracy": [
594
  0.01,
@@ -600,10 +1170,20 @@
600
  0.0,
601
  0.0,
602
  0.0,
 
 
 
 
 
 
 
 
 
 
603
  0.0
604
  ]
605
  },
606
- "final_accuracy": 0.8683333333333333,
607
  "sft_eval": {
608
  "config": {
609
  "ops": "add_sub",
@@ -614,7 +1194,7 @@
614
  },
615
  "splits": {
616
  "add_S0": {
617
- "full_accuracy": 0.98,
618
  "n_examples": 100,
619
  "per_subtask": {
620
  "SA": {
@@ -622,17 +1202,17 @@
622
  "count": 605
623
  },
624
  "SS": {
625
- "accuracy": 0.9789473684210527,
626
  "count": 95
627
  }
628
  }
629
  },
630
  "add_S1": {
631
- "full_accuracy": 0.99,
632
  "n_examples": 100,
633
  "per_subtask": {
634
  "SA": {
635
- "accuracy": 0.9950980392156863,
636
  "count": 204
637
  },
638
  "SC": {
@@ -650,7 +1230,7 @@
650
  }
651
  },
652
  "add_S2": {
653
- "full_accuracy": 0.85,
654
  "n_examples": 100,
655
  "per_subtask": {
656
  "SA": {
@@ -658,33 +1238,33 @@
658
  "count": 163
659
  },
660
  "SC": {
661
- "accuracy": 0.9615384615384616,
662
  "count": 130
663
  },
664
  "SS": {
665
- "accuracy": 0.9655172413793104,
666
  "count": 87
667
  },
668
  "UC": {
669
- "accuracy": 0.9655172413793104,
670
  "count": 203
671
  },
672
  "US": {
673
- "accuracy": 0.9914529914529915,
674
  "count": 117
675
  }
676
  }
677
  },
678
  "add_S3": {
679
- "full_accuracy": 0.5,
680
  "n_examples": 100,
681
  "per_subtask": {
682
  "SA": {
683
- "accuracy": 0.9834710743801653,
684
  "count": 121
685
  },
686
  "SC": {
687
- "accuracy": 0.9752066115702479,
688
  "count": 121
689
  },
690
  "SS": {
@@ -692,17 +1272,17 @@
692
  "count": 49
693
  },
694
  "UC": {
695
- "accuracy": 0.7580645161290323,
696
  "count": 186
697
  },
698
  "US": {
699
- "accuracy": 0.9730941704035875,
700
  "count": 223
701
  }
702
  }
703
  },
704
  "add_S4": {
705
- "full_accuracy": 0.37,
706
  "n_examples": 100,
707
  "per_subtask": {
708
  "SA": {
@@ -710,7 +1290,7 @@
710
  "count": 104
711
  },
712
  "SC": {
713
- "accuracy": 0.9811320754716981,
714
  "count": 106
715
  },
716
  "SS": {
@@ -718,17 +1298,17 @@
718
  "count": 23
719
  },
720
  "UC": {
721
- "accuracy": 0.66875,
722
  "count": 160
723
  },
724
  "US": {
725
- "accuracy": 0.8045602605863192,
726
  "count": 307
727
  }
728
  }
729
  },
730
  "add_S5": {
731
- "full_accuracy": 0.18,
732
  "n_examples": 100,
733
  "per_subtask": {
734
  "SA": {
@@ -740,17 +1320,17 @@
740
  "count": 100
741
  },
742
  "UC": {
743
- "accuracy": 0.3,
744
  "count": 100
745
  },
746
  "US": {
747
- "accuracy": 0.605,
748
  "count": 400
749
  }
750
  }
751
  },
752
  "add_S6": {
753
- "full_accuracy": 0.38,
754
  "n_examples": 100,
755
  "per_subtask": {
756
  "SC": {
@@ -758,25 +1338,25 @@
758
  "count": 100
759
  },
760
  "UC": {
761
- "accuracy": 0.53,
762
  "count": 100
763
  },
764
  "US": {
765
- "accuracy": 0.644,
766
  "count": 500
767
  }
768
  }
769
  },
770
  "add_random": {
771
- "full_accuracy": 0.925,
772
  "n_examples": 200,
773
  "per_subtask": {
774
  "SA": {
775
- "accuracy": 0.9910514541387024,
776
  "count": 447
777
  },
778
  "SC": {
779
- "accuracy": 0.984375,
780
  "count": 320
781
  },
782
  "SS": {
@@ -784,17 +1364,17 @@
784
  "count": 56
785
  },
786
  "UC": {
787
- "accuracy": 0.9886578449905482,
788
  "count": 529
789
  },
790
  "US": {
791
- "accuracy": 0.9791666666666666,
792
  "count": 48
793
  }
794
  }
795
  },
796
  "add_C3": {
797
- "full_accuracy": 0.81,
798
  "n_examples": 100,
799
  "per_subtask": {
800
  "SA": {
@@ -806,17 +1386,17 @@
806
  "count": 100
807
  },
808
  "UC": {
809
- "accuracy": 0.9067357512953368,
810
  "count": 193
811
  },
812
  "US": {
813
- "accuracy": 0.9626168224299065,
814
  "count": 107
815
  }
816
  }
817
  },
818
  "add_C4": {
819
- "full_accuracy": 0.75,
820
  "n_examples": 100,
821
  "per_subtask": {
822
  "SA": {
@@ -828,17 +1408,17 @@
828
  "count": 100
829
  },
830
  "UC": {
831
- "accuracy": 0.90625,
832
  "count": 256
833
  },
834
  "US": {
835
- "accuracy": 0.9305555555555556,
836
  "count": 144
837
  }
838
  }
839
  },
840
  "add_C5": {
841
- "full_accuracy": 0.68,
842
  "n_examples": 100,
843
  "per_subtask": {
844
  "SA": {
@@ -850,17 +1430,17 @@
850
  "count": 100
851
  },
852
  "UC": {
853
- "accuracy": 0.9084967320261438,
854
  "count": 306
855
  },
856
  "US": {
857
- "accuracy": 0.8556701030927835,
858
  "count": 194
859
  }
860
  }
861
  },
862
  "add_C6": {
863
- "full_accuracy": 0.76,
864
  "n_examples": 100,
865
  "per_subtask": {
866
  "SC": {
@@ -868,21 +1448,21 @@
868
  "count": 100
869
  },
870
  "UC": {
871
- "accuracy": 0.9344262295081968,
872
  "count": 366
873
  },
874
  "US": {
875
- "accuracy": 0.9700854700854701,
876
  "count": 234
877
  }
878
  }
879
  },
880
  "sub_M0": {
881
- "full_accuracy": 1.0,
882
  "n_examples": 100,
883
  "per_subtask": {
884
  "MD": {
885
- "accuracy": 1.0,
886
  "count": 601
887
  },
888
  "ME": {
@@ -892,29 +1472,29 @@
892
  }
893
  },
894
  "sub_M1": {
895
- "full_accuracy": 0.94,
896
  "n_examples": 100,
897
  "per_subtask": {
898
  "MD": {
899
- "accuracy": 0.996415770609319,
900
  "count": 279
901
  },
902
  "MB": {
903
- "accuracy": 0.9862068965517241,
904
  "count": 145
905
  },
906
  "ME": {
907
- "accuracy": 0.9583333333333334,
908
  "count": 24
909
  },
910
  "UB": {
911
- "accuracy": 0.9920634920634921,
912
  "count": 252
913
  }
914
  }
915
  },
916
  "sub_M2": {
917
- "full_accuracy": 0.9,
918
  "n_examples": 100,
919
  "per_subtask": {
920
  "MD": {
@@ -922,7 +1502,7 @@
922
  "count": 213
923
  },
924
  "MB": {
925
- "accuracy": 0.9911504424778761,
926
  "count": 113
927
  },
928
  "ME": {
@@ -930,7 +1510,7 @@
930
  "count": 85
931
  },
932
  "UB": {
933
- "accuracy": 0.9502762430939227,
934
  "count": 181
935
  },
936
  "UD": {
@@ -940,7 +1520,7 @@
940
  }
941
  },
942
  "sub_M3": {
943
- "full_accuracy": 0.19,
944
  "n_examples": 100,
945
  "per_subtask": {
946
  "MD": {
@@ -956,17 +1536,17 @@
956
  "count": 56
957
  },
958
  "UB": {
959
- "accuracy": 0.4697986577181208,
960
  "count": 149
961
  },
962
  "UD": {
963
- "accuracy": 0.9765258215962441,
964
  "count": 213
965
  }
966
  }
967
  },
968
  "sub_M4": {
969
- "full_accuracy": 0.01,
970
  "n_examples": 100,
971
  "per_subtask": {
972
  "MD": {
@@ -978,17 +1558,17 @@
978
  "count": 100
979
  },
980
  "UB": {
981
- "accuracy": 0.1,
982
  "count": 100
983
  },
984
  "UD": {
985
- "accuracy": 0.57,
986
  "count": 300
987
  }
988
  }
989
  },
990
  "sub_M5": {
991
- "full_accuracy": 0.0,
992
  "n_examples": 100,
993
  "per_subtask": {
994
  "MD": {
@@ -1000,17 +1580,17 @@
1000
  "count": 100
1001
  },
1002
  "UB": {
1003
- "accuracy": 0.02,
1004
  "count": 100
1005
  },
1006
  "UD": {
1007
- "accuracy": 0.435,
1008
  "count": 400
1009
  }
1010
  }
1011
  },
1012
  "sub_random": {
1013
- "full_accuracy": 0.96,
1014
  "n_examples": 200,
1015
  "per_subtask": {
1016
  "MD": {
@@ -1026,7 +1606,7 @@
1026
  "count": 53
1027
  },
1028
  "UB": {
1029
- "accuracy": 0.9886104783599089,
1030
  "count": 439
1031
  },
1032
  "UD": {
@@ -1036,11 +1616,11 @@
1036
  }
1037
  },
1038
  "sub_B3": {
1039
- "full_accuracy": 0.77,
1040
  "n_examples": 100,
1041
  "per_subtask": {
1042
  "MD": {
1043
- "accuracy": 1.0,
1044
  "count": 300
1045
  },
1046
  "MB": {
@@ -1048,17 +1628,17 @@
1048
  "count": 100
1049
  },
1050
  "UB": {
1051
- "accuracy": 0.883248730964467,
1052
  "count": 197
1053
  },
1054
  "UD": {
1055
- "accuracy": 0.9805825242718447,
1056
  "count": 103
1057
  }
1058
  }
1059
  },
1060
  "sub_B4": {
1061
- "full_accuracy": 0.6,
1062
  "n_examples": 100,
1063
  "per_subtask": {
1064
  "MD": {
@@ -1070,17 +1650,17 @@
1070
  "count": 100
1071
  },
1072
  "UB": {
1073
- "accuracy": 0.8380566801619433,
1074
  "count": 247
1075
  },
1076
  "UD": {
1077
- "accuracy": 0.8758169934640523,
1078
  "count": 153
1079
  }
1080
  }
1081
  },
1082
  "sub_B5": {
1083
- "full_accuracy": 0.5,
1084
  "n_examples": 100,
1085
  "per_subtask": {
1086
  "MD": {
@@ -1092,18 +1672,18 @@
1092
  "count": 100
1093
  },
1094
  "UB": {
1095
- "accuracy": 0.8355704697986577,
1096
  "count": 298
1097
  },
1098
  "UD": {
1099
- "accuracy": 0.8663366336633663,
1100
  "count": 202
1101
  }
1102
  }
1103
  }
1104
  },
1105
  "summary": {
1106
- "overall_accuracy": 0.66375,
1107
  "total_examples": 2400,
1108
  "n_splits": 22
1109
  }
@@ -1118,11 +1698,11 @@
1118
  },
1119
  "splits": {
1120
  "add_S0": {
1121
- "full_accuracy": 0.99,
1122
  "n_examples": 100,
1123
  "per_subtask": {
1124
  "SA": {
1125
- "accuracy": 0.9983471074380166,
1126
  "count": 605
1127
  },
1128
  "SS": {
@@ -1154,7 +1734,7 @@
1154
  }
1155
  },
1156
  "add_S2": {
1157
- "full_accuracy": 0.99,
1158
  "n_examples": 100,
1159
  "per_subtask": {
1160
  "SA": {
@@ -1170,7 +1750,7 @@
1170
  "count": 87
1171
  },
1172
  "UC": {
1173
- "accuracy": 0.9950738916256158,
1174
  "count": 203
1175
  },
1176
  "US": {
@@ -1180,7 +1760,7 @@
1180
  }
1181
  },
1182
  "add_S3": {
1183
- "full_accuracy": 0.98,
1184
  "n_examples": 100,
1185
  "per_subtask": {
1186
  "SA": {
@@ -1196,7 +1776,7 @@
1196
  "count": 49
1197
  },
1198
  "UC": {
1199
- "accuracy": 0.989247311827957,
1200
  "count": 186
1201
  },
1202
  "US": {
@@ -1206,7 +1786,7 @@
1206
  }
1207
  },
1208
  "add_S4": {
1209
- "full_accuracy": 0.86,
1210
  "n_examples": 100,
1211
  "per_subtask": {
1212
  "SA": {
@@ -1222,17 +1802,17 @@
1222
  "count": 23
1223
  },
1224
  "UC": {
1225
- "accuracy": 0.9125,
1226
  "count": 160
1227
  },
1228
  "US": {
1229
- "accuracy": 0.9869706840390879,
1230
  "count": 307
1231
  }
1232
  }
1233
  },
1234
  "add_S5": {
1235
- "full_accuracy": 0.35,
1236
  "n_examples": 100,
1237
  "per_subtask": {
1238
  "SA": {
@@ -1244,17 +1824,17 @@
1244
  "count": 100
1245
  },
1246
  "UC": {
1247
- "accuracy": 0.37,
1248
  "count": 100
1249
  },
1250
  "US": {
1251
- "accuracy": 0.8925,
1252
  "count": 400
1253
  }
1254
  }
1255
  },
1256
  "add_S6": {
1257
- "full_accuracy": 0.54,
1258
  "n_examples": 100,
1259
  "per_subtask": {
1260
  "SC": {
@@ -1262,25 +1842,25 @@
1262
  "count": 100
1263
  },
1264
  "UC": {
1265
- "accuracy": 0.55,
1266
  "count": 100
1267
  },
1268
  "US": {
1269
- "accuracy": 0.882,
1270
  "count": 500
1271
  }
1272
  }
1273
  },
1274
  "add_random": {
1275
- "full_accuracy": 0.98,
1276
  "n_examples": 200,
1277
  "per_subtask": {
1278
  "SA": {
1279
- "accuracy": 0.9932885906040269,
1280
  "count": 447
1281
  },
1282
  "SC": {
1283
- "accuracy": 0.996875,
1284
  "count": 320
1285
  },
1286
  "SS": {
@@ -1298,7 +1878,7 @@
1298
  }
1299
  },
1300
  "add_C3": {
1301
- "full_accuracy": 0.95,
1302
  "n_examples": 100,
1303
  "per_subtask": {
1304
  "SA": {
@@ -1310,7 +1890,7 @@
1310
  "count": 100
1311
  },
1312
  "UC": {
1313
- "accuracy": 0.9740932642487047,
1314
  "count": 193
1315
  },
1316
  "US": {
@@ -1320,7 +1900,7 @@
1320
  }
1321
  },
1322
  "add_C4": {
1323
- "full_accuracy": 0.96,
1324
  "n_examples": 100,
1325
  "per_subtask": {
1326
  "SA": {
@@ -1332,17 +1912,17 @@
1332
  "count": 100
1333
  },
1334
  "UC": {
1335
- "accuracy": 0.98828125,
1336
  "count": 256
1337
  },
1338
  "US": {
1339
- "accuracy": 0.9861111111111112,
1340
  "count": 144
1341
  }
1342
  }
1343
  },
1344
  "add_C5": {
1345
- "full_accuracy": 0.89,
1346
  "n_examples": 100,
1347
  "per_subtask": {
1348
  "SA": {
@@ -1354,17 +1934,17 @@
1354
  "count": 100
1355
  },
1356
  "UC": {
1357
- "accuracy": 0.9705882352941176,
1358
  "count": 306
1359
  },
1360
  "US": {
1361
- "accuracy": 0.9742268041237113,
1362
  "count": 194
1363
  }
1364
  }
1365
  },
1366
  "add_C6": {
1367
- "full_accuracy": 0.96,
1368
  "n_examples": 100,
1369
  "per_subtask": {
1370
  "SC": {
@@ -1372,11 +1952,11 @@
1372
  "count": 100
1373
  },
1374
  "UC": {
1375
- "accuracy": 0.9890710382513661,
1376
  "count": 366
1377
  },
1378
  "US": {
1379
- "accuracy": 0.9957264957264957,
1380
  "count": 234
1381
  }
1382
  }
@@ -1444,7 +2024,7 @@
1444
  }
1445
  },
1446
  "sub_M3": {
1447
- "full_accuracy": 0.95,
1448
  "n_examples": 100,
1449
  "per_subtask": {
1450
  "MD": {
@@ -1460,17 +2040,17 @@
1460
  "count": 56
1461
  },
1462
  "UB": {
1463
- "accuracy": 0.9664429530201343,
1464
  "count": 149
1465
  },
1466
  "UD": {
1467
- "accuracy": 0.9953051643192489,
1468
  "count": 213
1469
  }
1470
  }
1471
  },
1472
  "sub_M4": {
1473
- "full_accuracy": 0.55,
1474
  "n_examples": 100,
1475
  "per_subtask": {
1476
  "MD": {
@@ -1482,17 +2062,17 @@
1482
  "count": 100
1483
  },
1484
  "UB": {
1485
- "accuracy": 0.63,
1486
  "count": 100
1487
  },
1488
  "UD": {
1489
- "accuracy": 0.95,
1490
  "count": 300
1491
  }
1492
  }
1493
  },
1494
  "sub_M5": {
1495
- "full_accuracy": 0.11,
1496
  "n_examples": 100,
1497
  "per_subtask": {
1498
  "MD": {
@@ -1504,21 +2084,21 @@
1504
  "count": 100
1505
  },
1506
  "UB": {
1507
- "accuracy": 0.11,
1508
  "count": 100
1509
  },
1510
  "UD": {
1511
- "accuracy": 0.895,
1512
  "count": 400
1513
  }
1514
  }
1515
  },
1516
  "sub_random": {
1517
- "full_accuracy": 1.0,
1518
  "n_examples": 200,
1519
  "per_subtask": {
1520
  "MD": {
1521
- "accuracy": 1.0,
1522
  "count": 600
1523
  },
1524
  "MB": {
@@ -1540,7 +2120,7 @@
1540
  }
1541
  },
1542
  "sub_B3": {
1543
- "full_accuracy": 0.95,
1544
  "n_examples": 100,
1545
  "per_subtask": {
1546
  "MD": {
@@ -1552,17 +2132,17 @@
1552
  "count": 100
1553
  },
1554
  "UB": {
1555
- "accuracy": 0.9746192893401016,
1556
  "count": 197
1557
  },
1558
  "UD": {
1559
- "accuracy": 0.9902912621359223,
1560
  "count": 103
1561
  }
1562
  }
1563
  },
1564
  "sub_B4": {
1565
- "full_accuracy": 0.95,
1566
  "n_examples": 100,
1567
  "per_subtask": {
1568
  "MD": {
@@ -1574,17 +2154,17 @@
1574
  "count": 100
1575
  },
1576
  "UB": {
1577
- "accuracy": 0.9838056680161943,
1578
  "count": 247
1579
  },
1580
  "UD": {
1581
- "accuracy": 0.9803921568627451,
1582
  "count": 153
1583
  }
1584
  }
1585
  },
1586
  "sub_B5": {
1587
- "full_accuracy": 0.9,
1588
  "n_examples": 100,
1589
  "per_subtask": {
1590
  "MD": {
@@ -1596,22 +2176,22 @@
1596
  "count": 100
1597
  },
1598
  "UB": {
1599
- "accuracy": 0.9664429530201343,
1600
  "count": 298
1601
  },
1602
  "UD": {
1603
- "accuracy": 0.9801980198019802,
1604
  "count": 202
1605
  }
1606
  }
1607
  }
1608
  },
1609
  "summary": {
1610
- "overall_accuracy": 0.8683333333333333,
1611
  "total_examples": 2400,
1612
  "n_splits": 22
1613
  }
1614
  },
1615
- "sorl_overall_accuracy": 0.8683333333333333,
1616
- "sft_overall_accuracy": 0.66375
1617
  }
 
70
  3719,
71
  3769,
72
  3819,
73
+ 3869,
74
+ 3960,
75
+ 4010,
76
+ 4060,
77
+ 4110,
78
+ 4160,
79
+ 4210,
80
+ 4260,
81
+ 4351,
82
+ 4401,
83
+ 4451,
84
+ 4501,
85
+ 4551,
86
+ 4601,
87
+ 4651,
88
+ 4742,
89
+ 4792,
90
+ 4842,
91
+ 4892,
92
+ 4942,
93
+ 4992,
94
+ 5042,
95
+ 5133,
96
+ 5183,
97
+ 5233,
98
+ 5283,
99
+ 5333,
100
+ 5383,
101
+ 5433,
102
+ 5524,
103
+ 5574,
104
+ 5624,
105
+ 5674,
106
+ 5724,
107
+ 5774,
108
+ 5824,
109
+ 5915,
110
+ 5965,
111
+ 6015,
112
+ 6065,
113
+ 6115,
114
+ 6165,
115
+ 6215,
116
+ 6306,
117
+ 6356,
118
+ 6406,
119
+ 6456,
120
+ 6506,
121
+ 6556,
122
+ 6606,
123
+ 6697,
124
+ 6747,
125
+ 6797,
126
+ 6847,
127
+ 6897,
128
+ 6947,
129
+ 6997,
130
+ 7088,
131
+ 7138,
132
+ 7188,
133
+ 7238,
134
+ 7288,
135
+ 7338,
136
+ 7388,
137
+ 7479,
138
+ 7529,
139
+ 7579,
140
+ 7629,
141
+ 7679,
142
+ 7729,
143
+ 7779
144
  ],
145
  "loss": [
146
+ 2.775237560272217,
147
+ 7.547610282897949,
148
+ 4.619787216186523,
149
+ 3.3733177185058594,
150
+ 2.9553494453430176,
151
+ 2.9782869815826416,
152
+ 2.651740550994873,
153
+ -4.420749664306641,
154
+ -4.0340046882629395,
155
+ -1.603489875793457,
156
+ -0.8856117725372314,
157
+ -0.3331648111343384,
158
+ -0.7166460752487183,
159
+ -0.9935495853424072,
160
+ -0.5348667502403259,
161
+ -0.2662088871002197,
162
+ -0.36422669887542725,
163
+ -0.3612262010574341,
164
+ 0.24359634518623352,
165
+ -0.1940872073173523,
166
+ -0.2658434510231018,
167
+ -0.4549145996570587,
168
+ -0.7099339365959167,
169
+ -0.3537546396255493,
170
+ -0.6272122859954834,
171
+ -0.6276133060455322,
172
+ -0.7034407258033752,
173
+ -0.8803790807723999,
174
+ -0.6200196743011475,
175
+ -1.0295164585113525,
176
+ -0.5204843282699585,
177
+ -0.4025934636592865,
178
+ -0.6784000992774963,
179
+ -0.4574926495552063,
180
+ -0.6503784656524658,
181
+ -0.21740400791168213,
182
+ -0.3695314824581146,
183
+ -0.43050533533096313,
184
+ -0.44321656227111816,
185
+ -0.45792967081069946,
186
+ -0.2178489714860916,
187
+ -0.6612284183502197,
188
+ -0.36298084259033203,
189
+ -0.3561042249202728,
190
+ -0.5411268472671509,
191
+ -0.15997686982154846,
192
+ -0.2238040417432785,
193
+ -0.7527804374694824,
194
+ -0.25649505853652954,
195
+ -0.6289592981338501,
196
+ -0.49410122632980347,
197
+ -0.18331468105316162,
198
+ -0.3199687898159027,
199
+ -0.2754635810852051,
200
+ -0.6548870205879211,
201
+ -0.37214863300323486,
202
+ 0.16730983555316925,
203
+ -0.25607019662857056,
204
+ -0.5547778010368347,
205
+ -0.23881055414676666,
206
+ -0.23869141936302185,
207
+ -0.3053569197654724,
208
+ -0.6120729446411133,
209
+ -0.17358490824699402,
210
+ -0.41464683413505554,
211
+ -0.3155255615711212,
212
+ -0.539973795413971,
213
+ -0.3112099766731262,
214
+ -0.529899001121521,
215
+ -0.22048673033714294,
216
+ 0.12720385193824768,
217
+ -0.1732008457183838,
218
+ -0.09826218336820602,
219
+ -0.09295424073934555,
220
+ -0.1722036600112915,
221
+ -0.32555434107780457,
222
+ -0.06960314512252808,
223
+ -0.24585965275764465,
224
+ -0.2691691815853119,
225
+ -0.07430160790681839,
226
+ -0.12875692546367645,
227
+ -0.49890026450157166,
228
+ -0.03004104644060135,
229
+ -0.27030062675476074,
230
+ -0.4276649057865143,
231
+ -0.24829161167144775,
232
+ -0.24454763531684875,
233
+ -0.20643892884254456,
234
+ -0.23065191507339478,
235
+ -0.15173403918743134,
236
+ -0.1288289725780487,
237
+ -0.5371881723403931,
238
+ -0.18638940155506134,
239
+ -0.39545783400535583,
240
+ -0.12707214057445526,
241
+ -0.10258357971906662,
242
+ -0.21331043541431427,
243
+ -0.3237987756729126,
244
+ -0.2645500898361206,
245
+ -0.24036605656147003,
246
+ -0.34151068329811096,
247
+ -0.11535628139972687,
248
+ -0.10533130913972855,
249
+ -0.17564454674720764,
250
+ -0.10551251471042633,
251
+ -0.18365180492401123,
252
+ -0.2547161281108856,
253
+ -0.08338633179664612,
254
+ -0.017869174480438232,
255
+ -0.07794929295778275,
256
+ -0.2373967319726944,
257
+ -0.07128022611141205,
258
+ -0.23635011911392212,
259
+ -0.13728651404380798,
260
+ -0.028313498944044113,
261
+ -0.14107288420200348,
262
+ -0.27062463760375977,
263
+ -0.08939403295516968,
264
+ -0.1667688935995102,
265
+ -0.16491536796092987,
266
+ -0.12269027531147003,
267
+ -0.36350658535957336,
268
+ -0.05644230544567108,
269
+ -0.010060478001832962,
270
+ -0.15089525282382965,
271
+ -0.08028542250394821,
272
+ -0.07134267687797546,
273
+ -0.25595834851264954,
274
+ -0.17725200951099396,
275
+ -0.057140905410051346,
276
+ -0.04304641857743263,
277
+ -0.11805126816034317,
278
+ -0.0738956555724144,
279
+ -0.028574828058481216,
280
+ -0.032799072563648224,
281
+ -0.023946844041347504,
282
+ -0.07897244393825531,
283
+ -0.038176145404577255,
284
+ -0.0022059171460568905,
285
+ -0.04675590991973877
286
  ],
287
  "base_loss": [
288
+ 9.28781795501709,
289
+ 6.408207416534424,
290
+ 3.954561471939087,
291
+ 2.0575625896453857,
292
+ 1.966015100479126,
293
+ 1.913878321647644,
294
+ 1.828439712524414,
295
+ 1.8617305755615234,
296
+ 1.431872010231018,
297
+ 0.9288164377212524,
298
+ 0.7166451215744019,
299
+ 0.5394684076309204,
300
+ 0.5172553658485413,
301
+ 0.5257603526115417,
302
+ 0.3500528931617737,
303
+ 0.2768792510032654,
304
+ 0.2930271327495575,
305
+ 0.28543582558631897,
306
+ 0.2047962099313736,
307
+ 0.21444563567638397,
308
+ 0.20364591479301453,
309
+ 0.1602955162525177,
310
+ 0.17975081503391266,
311
+ 0.10734299570322037,
312
+ 0.1685129553079605,
313
+ 0.17823576927185059,
314
+ 0.1294601410627365,
315
+ 0.1530643254518509,
316
+ 0.12994277477264404,
317
+ 0.1753716617822647,
318
+ 0.098540760576725,
319
+ 0.13171425461769104,
320
+ 0.12440256029367447,
321
+ 0.09750843793153763,
322
+ 0.10259241610765457,
323
+ 0.08516658842563629,
324
+ 0.0938568040728569,
325
+ 0.08442933112382889,
326
+ 0.08345118910074234,
327
+ 0.07648413628339767,
328
+ 0.08328725397586823,
329
+ 0.10315794497728348,
330
+ 0.07307988405227661,
331
+ 0.06590916961431503,
332
+ 0.08993113040924072,
333
+ 0.05706944689154625,
334
+ 0.06592441350221634,
335
+ 0.09816405922174454,
336
+ 0.06872431188821793,
337
+ 0.08560836315155029,
338
+ 0.08522152900695801,
339
+ 0.03932935744524002,
340
+ 0.06628448516130447,
341
+ 0.054641298949718475,
342
+ 0.08359350264072418,
343
+ 0.06020509451627731,
344
+ 0.08234674483537674,
345
+ 0.04314882308244705,
346
+ 0.07642664015293121,
347
+ 0.0344601608812809,
348
+ 0.048634521663188934,
349
+ 0.04398999363183975,
350
+ 0.08435793966054916,
351
+ 0.04250624403357506,
352
+ 0.05525727942585945,
353
+ 0.06740889698266983,
354
+ 0.06672954559326172,
355
+ 0.06832120567560196,
356
+ 0.06656587868928909,
357
+ 0.06475470215082169,
358
+ 0.02944616600871086,
359
+ 0.02783912792801857,
360
+ 0.02230864204466343,
361
+ 0.01960669830441475,
362
+ 0.03943309187889099,
363
+ 0.04117203876376152,
364
+ 0.035829126834869385,
365
+ 0.03276657313108444,
366
+ 0.034448444843292236,
367
+ 0.05022916570305824,
368
+ 0.024941788986325264,
369
+ 0.06100919470191002,
370
+ 0.01847134344279766,
371
+ 0.04928785562515259,
372
+ 0.05794186145067215,
373
+ 0.03713752701878548,
374
+ 0.04499039426445961,
375
+ 0.028732936829328537,
376
+ 0.032221321016550064,
377
+ 0.020824583247303963,
378
+ 0.028740141540765762,
379
+ 0.06568349152803421,
380
+ 0.024854103103280067,
381
+ 0.04695362597703934,
382
+ 0.026407036930322647,
383
+ 0.016328101977705956,
384
+ 0.028396395966410637,
385
+ 0.0408906415104866,
386
+ 0.03339320793747902,
387
+ 0.04666588082909584,
388
+ 0.04415905103087425,
389
+ 0.018016483634710312,
390
+ 0.014827796258032322,
391
+ 0.023171423003077507,
392
+ 0.03068389557301998,
393
+ 0.022415442392230034,
394
+ 0.03334091231226921,
395
+ 0.010826569981873035,
396
+ 0.004939792212098837,
397
+ 0.011606397107243538,
398
+ 0.029728740453720093,
399
+ 0.011070964857935905,
400
+ 0.027625424787402153,
401
+ 0.01636265404522419,
402
+ 0.03281437233090401,
403
+ 0.01786724664270878,
404
+ 0.032829318195581436,
405
+ 0.011172984726727009,
406
+ 0.020385416224598885,
407
+ 0.01945609226822853,
408
+ 0.015599744394421577,
409
+ 0.0463777594268322,
410
+ 0.007090721745043993,
411
+ 0.002301769098266959,
412
+ 0.018458489328622818,
413
+ 0.010156680829823017,
414
+ 0.008803258650004864,
415
+ 0.029838016256690025,
416
+ 0.02362155355513096,
417
+ 0.007222625892609358,
418
+ 0.005627515260130167,
419
+ 0.013934285379946232,
420
+ 0.00901535339653492,
421
+ 0.003945329692214727,
422
+ 0.004317434038966894,
423
+ 0.004925424698740244,
424
+ 0.009535633958876133,
425
+ 0.005236275494098663,
426
+ 0.000816056621260941,
427
+ 0.006017677485942841
428
  ],
429
  "info_loss": [
430
+ -1.4931654930114746,
431
+ -0.3287515640258789,
432
+ -0.15673398971557617,
433
+ -0.060143470764160156,
434
+ -0.08957064151763916,
435
+ -0.08109843730926514,
436
+ -0.10492265224456787,
437
+ -0.8159008026123047,
438
+ -0.7340517640113831,
439
+ -0.436699241399765,
440
+ -0.3305722177028656,
441
+ -0.2345598042011261,
442
+ -0.26151734590530396,
443
+ -0.2701610028743744,
444
+ -0.1915142834186554,
445
+ -0.1484409123659134,
446
+ -0.1524379551410675,
447
+ -0.13985788822174072,
448
+ -0.07239346206188202,
449
+ -0.11426207423210144,
450
+ -0.10004279017448425,
451
+ -0.08696124702692032,
452
+ -0.10689475387334824,
453
+ -0.06413023173809052,
454
+ -0.09764015674591064,
455
+ -0.09936109930276871,
456
+ -0.10012608021497726,
457
+ -0.11616788059473038,
458
+ -0.08879338204860687,
459
+ -0.13839909434318542,
460
+ -0.075227752327919,
461
+ -0.06567079573869705,
462
+ -0.09227447211742401,
463
+ -0.06739025563001633,
464
+ -0.08406399190425873,
465
+ -0.04091668874025345,
466
+ -0.05387166142463684,
467
+ -0.05921587347984314,
468
+ -0.060556016862392426,
469
+ -0.06337527930736542,
470
+ -0.037323929369449615,
471
+ -0.08630433678627014,
472
+ -0.04977574944496155,
473
+ -0.049126606434583664,
474
+ -0.07182436436414719,
475
+ -0.027198320254683495,
476
+ -0.034297745674848557,
477
+ -0.09243270754814148,
478
+ -0.03931873291730881,
479
+ -0.0762982964515686,
480
+ -0.06489811837673187,
481
+ -0.027625108137726784,
482
+ -0.048267342150211334,
483
+ -0.041161395609378815,
484
+ -0.08079905807971954,
485
+ -0.04790109768509865,
486
+ 0.0011650696396827698,
487
+ -0.03641394525766373,
488
+ -0.06875798851251602,
489
+ -0.03174798563122749,
490
+ -0.033739905804395676,
491
+ -0.038479916751384735,
492
+ -0.07334434241056442,
493
+ -0.025829708203673363,
494
+ -0.051409944891929626,
495
+ -0.04240108281373978,
496
+ -0.06412427872419357,
497
+ -0.0410870797932148,
498
+ -0.06333422660827637,
499
+ -0.03284423425793648,
500
+ 0.00417708232998848,
501
+ -0.024818118661642075,
502
+ -0.017061999067664146,
503
+ -0.016985513269901276,
504
+ -0.02596151828765869,
505
+ -0.04014855995774269,
506
+ -0.015163155272603035,
507
+ -0.03072977066040039,
508
+ -0.0335262157022953,
509
+ -0.016495104879140854,
510
+ -0.017648227512836456,
511
+ -0.05817177891731262,
512
+ -0.009896370582282543,
513
+ -0.03396162390708923,
514
+ -0.052463896572589874,
515
+ -0.031708974391222,
516
+ -0.031061530113220215,
517
+ -0.02669505402445793,
518
+ -0.02900986559689045,
519
+ -0.019973134621977806,
520
+ -0.018899552524089813,
521
+ -0.06360407918691635,
522
+ -0.024449247866868973,
523
+ -0.046435195952653885,
524
+ -0.01775706559419632,
525
+ -0.014864739961922169,
526
+ -0.025742633268237114,
527
+ -0.03865946829319,
528
+ -0.03197530657052994,
529
+ -0.031122641637921333,
530
+ -0.04157198593020439,
531
+ -0.01486174762248993,
532
+ -0.014675622805953026,
533
+ -0.02121778391301632,
534
+ -0.014983661472797394,
535
+ -0.022015109658241272,
536
+ -0.03083636984229088,
537
+ -0.010758324526250362,
538
+ -0.0045216502621769905,
539
+ -0.01022826973348856,
540
+ -0.028052691370248795,
541
+ -0.010628215968608856,
542
+ -0.027519239112734795,
543
+ -0.016260437667369843,
544
+ -0.006960311904549599,
545
+ -0.017028087750077248,
546
+ -0.03268968313932419,
547
+ -0.011114806868135929,
548
+ -0.020086344331502914,
549
+ -0.019388578832149506,
550
+ -0.015157317742705345,
551
+ -0.04190883785486221,
552
+ -0.007044406142085791,
553
+ -0.0022618311922997236,
554
+ -0.01841524988412857,
555
+ -0.010110707953572273,
556
+ -0.008758192881941795,
557
+ -0.029799800366163254,
558
+ -0.021984053775668144,
559
+ -0.007137994281947613,
560
+ -0.005599978379905224,
561
+ -0.013903001323342323,
562
+ -0.00883825309574604,
563
+ -0.0038955863565206528,
564
+ -0.004232798703014851,
565
+ -0.004899953026324511,
566
+ -0.009251396171748638,
567
+ -0.0052089691162109375,
568
+ -0.0007940736250020564,
569
+ -0.005982651840895414
570
  ],
571
  "abs_loss": [
572
+ 2.287022590637207,
573
+ 2.0916876792907715,
574
+ 1.8786654472351074,
575
+ 1.8568774461746216,
576
+ 1.8498092889785767,
577
+ 1.8190327882766724,
578
+ 1.8224343061447144,
579
+ 1.8426017761230469,
580
+ 1.8160287141799927,
581
+ 1.5422773361206055,
582
+ 1.1873483657836914,
583
+ 0.7351039052009583,
584
+ 0.7013891339302063,
585
+ 0.4904753267765045,
586
+ 0.39240923523902893,
587
+ 0.40737900137901306,
588
+ 0.32288554310798645,
589
+ 0.3128460645675659,
590
+ 0.32565391063690186,
591
+ 0.27980709075927734,
592
+ 0.23081153631210327,
593
+ 0.164260134100914,
594
+ 0.2405962496995926,
595
+ 0.20454931259155273,
596
+ 0.18585167825222015,
597
+ 0.1410851627588272,
598
+ 0.1597604751586914,
599
+ 0.13912147283554077,
600
+ 0.14329303801059723,
601
+ 0.1450016349554062,
602
+ 0.1277744472026825,
603
+ 0.144809290766716,
604
+ 0.11082559078931808,
605
+ 0.09187864512205124,
606
+ 0.08627257496118546,
607
+ 0.09445662051439285,
608
+ 0.09944355487823486,
609
+ 0.0844881609082222,
610
+ 0.06807096302509308,
611
+ 0.0774887427687645,
612
+ 0.05698530003428459,
613
+ 0.06316094845533371,
614
+ 0.05285605788230896,
615
+ 0.07560881972312927,
616
+ 0.05246293544769287,
617
+ 0.06453344970941544,
618
+ 0.04189883545041084,
619
+ 0.057164814323186874,
620
+ 0.06852755695581436,
621
+ 0.06694110482931137,
622
+ 0.05445839464664459,
623
+ 0.04003268480300903,
624
+ 0.04949067905545235,
625
+ 0.06563352793455124,
626
+ 0.05591483786702156,
627
+ 0.03960074856877327,
628
+ 0.050677765160799026,
629
+ 0.05222274735569954,
630
+ 0.044354382902383804,
631
+ 0.04689810797572136,
632
+ 0.04471694305539131,
633
+ 0.033734675496816635,
634
+ 0.03260963782668114,
635
+ 0.034053441137075424,
636
+ 0.03475390747189522,
637
+ 0.027156146243214607,
638
+ 0.03657558932900429,
639
+ 0.032038044184446335,
640
+ 0.04324740171432495,
641
+ 0.029938995838165283,
642
+ 0.03984285518527031,
643
+ 0.03961440548300743,
644
+ 0.017977066338062286,
645
+ 0.04167564585804939,
646
+ 0.030039796605706215,
647
+ 0.020334169268608093,
648
+ 0.027248024940490723,
649
+ 0.02041912078857422,
650
+ 0.013136718422174454,
651
+ 0.02858579158782959,
652
+ 0.020576881244778633,
653
+ 0.033657435327768326,
654
+ 0.02780439890921116,
655
+ 0.053980231285095215,
656
+ 0.025031788274645805,
657
+ 0.02547328732907772,
658
+ 0.02095186896622181,
659
+ 0.02673853375017643,
660
+ 0.022440383210778236,
661
+ 0.03095315955579281,
662
+ 0.016164809465408325,
663
+ 0.03688853606581688,
664
+ 0.03149260953068733,
665
+ 0.020118767395615578,
666
+ 0.015789978206157684,
667
+ 0.026521777734160423,
668
+ 0.025128299370408058,
669
+ 0.015657806769013405,
670
+ 0.020167004317045212,
671
+ 0.017777908593416214,
672
+ 0.020459117367863655,
673
+ 0.020475341007113457,
674
+ 0.017835237085819244,
675
+ 0.013575234450399876,
676
+ 0.016010567545890808,
677
+ 0.010493863373994827,
678
+ 0.009120851755142212,
679
+ 0.013193558901548386,
680
+ 0.012166949920356274,
681
+ 0.01292345393449068,
682
+ 0.018737776204943657,
683
+ 0.018054399639368057,
684
+ 0.006013364065438509,
685
+ 0.018393641337752342,
686
+ 0.008320348337292671,
687
+ 0.018457762897014618,
688
+ 0.014404929243028164,
689
+ 0.012825693935155869,
690
+ 0.012252169661223888,
691
+ 0.01264596451073885,
692
+ 0.008783203549683094,
693
+ 0.006435094866901636,
694
+ 0.009466757997870445,
695
+ 0.011133003048598766,
696
+ 0.008680086582899094,
697
+ 0.014923741109669209,
698
+ 0.005890031810849905,
699
+ 0.009024207480251789,
700
+ 0.008068445138633251,
701
+ 0.007878277450799942,
702
+ 0.007220780476927757,
703
+ 0.006559442728757858,
704
+ 0.009277471341192722,
705
+ 0.003974832594394684,
706
+ 0.008603575639426708,
707
+ 0.009288758039474487,
708
+ 0.013746594078838825,
709
+ 0.013656378723680973,
710
+ 0.008018574677407742,
711
+ 0.012013244442641735
712
  ],
713
  "zipf_loss": [
714
+ 8.190372467041016,
715
+ 4.21774959564209,
716
+ 2.044699192047119,
717
+ 1.7315019369125366,
718
+ 1.7000598907470703,
719
+ 1.6934897899627686,
720
+ 1.690284013748169,
721
+ 1.692267656326294,
722
+ 1.6930378675460815,
723
+ 1.6804585456848145,
724
+ 1.5847303867340088,
725
+ 1.3994543552398682,
726
+ 1.3111330270767212,
727
+ 1.1332526206970215,
728
+ 0.9909822344779968,
729
+ 0.900583028793335,
730
+ 0.8348370790481567,
731
+ 0.7206323146820068,
732
+ 0.730169415473938,
733
+ 0.7061072587966919,
734
+ 0.5078573822975159,
735
+ 0.2379763424396515,
736
+ 0.15520316362380981,
737
+ 0.15974974632263184,
738
+ 0.1620911955833435,
739
+ 0.17365333437919617,
740
+ 0.15238380432128906,
741
+ 0.11432318389415741,
742
+ 0.12364210188388824,
743
+ 0.16460275650024414,
744
+ 0.1204749196767807,
745
+ 0.10791924595832825,
746
+ 0.10885953158140182,
747
+ 0.10971362143754959,
748
+ 0.07904183119535446,
749
+ 0.09715063124895096,
750
+ 0.06538397073745728,
751
+ 0.06877526640892029,
752
+ 0.07208532094955444,
753
+ 0.09159012138843536,
754
+ 0.06640453636646271,
755
+ 0.09234096854925156,
756
+ 0.05641115829348564,
757
+ 0.06169179081916809,
758
+ 0.08193936944007874,
759
+ 0.048483528196811676,
760
+ 0.04905913397669792,
761
+ 0.06766603887081146,
762
+ 0.061115242540836334,
763
+ 0.04172118008136749,
764
+ 0.06421263515949249,
765
+ 0.04960378259420395,
766
+ 0.09147107601165771,
767
+ 0.07494569569826126,
768
+ 0.06391854584217072,
769
+ 0.04269716516137123,
770
+ 0.06824462115764618,
771
+ 0.059698160737752914,
772
+ 0.051940008997917175,
773
+ 0.03951932117342949,
774
+ 0.045601435005664825,
775
+ 0.032078806310892105,
776
+ 0.03375157713890076,
777
+ 0.03880057483911514,
778
+ 0.040719956159591675,
779
+ 0.03836074844002724,
780
+ 0.030881881713867188,
781
+ 0.028135811910033226,
782
+ 0.03255265951156616,
783
+ 0.040206994861364365,
784
+ 0.05200258642435074,
785
+ 0.04317977651953697,
786
+ 0.04825145751237869,
787
+ 0.05312662571668625,
788
+ 0.044974446296691895,
789
+ 0.03272582218050957,
790
+ 0.0434744693338871,
791
+ 0.026629559695720673,
792
+ 0.030330833047628403,
793
+ 0.03756170719861984,
794
+ 0.020725876092910767,
795
+ 0.018442610278725624,
796
+ 0.047670878469944,
797
+ 0.014629715122282505,
798
+ 0.036528993397951126,
799
+ 0.02911325916647911,
800
+ 0.01898207701742649,
801
+ 0.02910483255982399,
802
+ 0.024981416761875153,
803
+ 0.024077415466308594,
804
+ 0.029809938743710518,
805
+ 0.02948029898107052,
806
+ 0.03009970486164093,
807
+ 0.019928639754652977,
808
+ 0.022512484341859818,
809
+ 0.027083536610007286,
810
+ 0.013206664472818375,
811
+ 0.020339498296380043,
812
+ 0.01979305036365986,
813
+ 0.022416699677705765,
814
+ 0.028004225343465805,
815
+ 0.013197165913879871,
816
+ 0.02481359988451004,
817
+ 0.012004347518086433,
818
+ 0.01203914638608694,
819
+ 0.013034461066126823,
820
+ 0.01939459890127182,
821
+ 0.012050990015268326,
822
+ 0.021190840750932693,
823
+ 0.011434659361839294,
824
+ 0.011527659371495247,
825
+ 0.02212553098797798,
826
+ 0.010615499690175056,
827
+ 0.007115839049220085,
828
+ 0.007643207907676697,
829
+ 0.009494960308074951,
830
+ 0.022002404555678368,
831
+ 0.0092984763905406,
832
+ 0.012483924627304077,
833
+ 0.008249725215137005,
834
+ 0.012404844164848328,
835
+ 0.008560520596802235,
836
+ 0.00596435833722353,
837
+ 0.009142763912677765,
838
+ 0.013930749148130417,
839
+ 0.009172595106065273,
840
+ 0.006846996955573559,
841
+ 0.011299213394522667,
842
+ 0.018160128965973854,
843
+ 0.0062285917811095715,
844
+ 0.006603773683309555,
845
+ 0.006388517562299967,
846
+ 0.004543771501630545,
847
+ 0.006038222461938858,
848
+ 0.004351124167442322,
849
+ 0.019198384135961533,
850
+ 0.002631222829222679,
851
+ 0.007311630062758923,
852
+ 0.004116905387490988,
853
+ 0.00585160730406642
854
  ],
855
  "denoise_loss": [],
856
  "ortho_loss": [
857
+ 0.6088240742683411,
858
+ 0.33010056614875793,
859
+ 0.21411478519439697,
860
+ 0.15372449159622192,
861
+ 0.10954908281564713,
862
+ 0.10619203746318817,
863
+ 0.1071133241057396,
864
+ 0.17322593927383423,
865
+ 0.18734809756278992,
866
+ 0.21046631038188934,
867
+ 0.2544556260108948,
868
+ 0.2811073660850525,
869
+ 0.27933359146118164,
870
+ 0.30032429099082947,
871
+ 0.3181747794151306,
872
+ 0.3021242916584015,
873
+ 0.30070003867149353,
874
+ 0.3047161102294922,
875
+ 0.29878291487693787,
876
+ 0.294882595539093,
877
+ 0.3017619848251343,
878
+ 0.3246193826198578,
879
+ 0.3274792730808258,
880
+ 0.31516149640083313,
881
+ 0.31390562653541565,
882
+ 0.3176639974117279,
883
+ 0.30613771080970764,
884
+ 0.30903494358062744,
885
+ 0.29815226793289185,
886
+ 0.3041422963142395,
887
+ 0.30682864785194397,
888
+ 0.29478564858436584,
889
+ 0.285143107175827,
890
+ 0.28599533438682556,
891
+ 0.28144142031669617,
892
+ 0.2649855315685272,
893
+ 0.27516791224479675,
894
+ 0.2877526581287384,
895
+ 0.2752108573913574,
896
+ 0.2736209034919739,
897
+ 0.2612563371658325,
898
+ 0.2680782675743103,
899
+ 0.2876756489276886,
900
+ 0.31357342004776,
901
+ 0.30909889936447144,
902
+ 0.29838162660598755,
903
+ 0.2932455837726593,
904
+ 0.29046431183815,
905
+ 0.2770218253135681,
906
+ 0.2851712703704834,
907
+ 0.2728922963142395,
908
+ 0.2799801230430603,
909
+ 0.28310680389404297,
910
+ 0.27114954590797424,
911
+ 0.29482606053352356,
912
+ 0.309641033411026,
913
+ 0.3022473156452179,
914
+ 0.2983833849430084,
915
+ 0.2894889712333679,
916
+ 0.2985239624977112,
917
+ 0.29320472478866577,
918
+ 0.2976941168308258,
919
+ 0.3074931204319,
920
+ 0.2983621060848236,
921
+ 0.28685250878334045,
922
+ 0.28766006231307983,
923
+ 0.2761837840080261,
924
+ 0.27767765522003174,
925
+ 0.26966720819473267,
926
+ 0.2801859676837921,
927
+ 0.27615296840667725,
928
+ 0.2672549784183502,
929
+ 0.2738680839538574,
930
+ 0.28428056836128235,
931
+ 0.274615615606308,
932
+ 0.2717064321041107,
933
+ 0.27200043201446533,
934
+ 0.2775023281574249,
935
+ 0.28582414984703064,
936
+ 0.2739453613758087,
937
+ 0.28235429525375366,
938
+ 0.2763129472732544,
939
+ 0.28233736753463745,
940
+ 0.2881776988506317,
941
+ 0.277167946100235,
942
+ 0.2771058678627014,
943
+ 0.26995790004730225,
944
+ 0.2737855911254883,
945
+ 0.2769392430782318,
946
+ 0.27080678939819336,
947
+ 0.26978862285614014,
948
+ 0.2715676426887512,
949
+ 0.2671617865562439,
950
+ 0.26437824964523315,
951
+ 0.2694927155971527,
952
+ 0.28054067492485046,
953
+ 0.2770703136920929,
954
+ 0.2670004963874817,
955
+ 0.2709817886352539,
956
+ 0.26673585176467896,
957
+ 0.2748439311981201,
958
+ 0.2779378890991211,
959
+ 0.27402395009994507,
960
+ 0.27640703320503235,
961
+ 0.2766958177089691,
962
+ 0.27871349453926086,
963
+ 0.27661144733428955,
964
+ 0.26058056950569153,
965
+ 0.2604549825191498,
966
+ 0.25930461287498474,
967
+ 0.25678619742393494,
968
+ 0.252104252576828,
969
+ 0.24827271699905396,
970
+ 0.2491685301065445,
971
+ 0.24485431611537933,
972
+ 0.24885989725589752,
973
+ 0.2484455108642578,
974
+ 0.2500342130661011,
975
+ 0.254868745803833,
976
+ 0.25000831484794617,
977
+ 0.2509933412075043,
978
+ 0.2539915442466736,
979
+ 0.2499563843011856,
980
+ 0.25075626373291016,
981
+ 0.2544092535972595,
982
+ 0.25233811140060425,
983
+ 0.25777095556259155,
984
+ 0.25766149163246155,
985
+ 0.25799405574798584,
986
+ 0.259542852640152,
987
+ 0.2591084837913513,
988
+ 0.2588949501514435,
989
+ 0.2605401277542114,
990
+ 0.25861841440200806,
991
+ 0.2574929893016815,
992
+ 0.257936030626297,
993
+ 0.25708767771720886,
994
+ 0.2581881284713745,
995
+ 0.2574787139892578,
996
+ 0.2564684748649597
997
  ],
998
  "lr": [
999
+ 1.6752136752136756e-05,
1000
+ 3.384615384615385e-05,
1001
+ 5.094017094017095e-05,
1002
+ 6.803418803418804e-05,
1003
  8e-05,
1004
  8e-05,
1005
  8e-05,
 
1040
  8e-05,
1041
  8e-05,
1042
  8e-05,
1043
+ 8e-05,
1044
+ 8e-05,
1045
+ 8e-05,
1046
+ 8e-05,
1047
+ 8e-05,
1048
+ 8e-05,
1049
+ 8e-05,
1050
+ 8e-05,
1051
+ 8e-05,
1052
+ 8e-05,
1053
+ 8e-05,
1054
+ 8e-05,
1055
+ 8e-05,
1056
+ 8e-05,
1057
+ 8e-05,
1058
+ 8e-05,
1059
+ 8e-05,
1060
+ 8e-05,
1061
+ 8e-05,
1062
+ 8e-05,
1063
+ 8e-05,
1064
+ 8e-05,
1065
+ 8e-05,
1066
+ 8e-05,
1067
+ 8e-05,
1068
+ 8e-05,
1069
+ 8e-05,
1070
+ 8e-05,
1071
+ 8e-05,
1072
+ 8e-05,
1073
+ 8e-05,
1074
+ 8e-05,
1075
+ 8e-05,
1076
+ 8e-05,
1077
+ 8e-05,
1078
+ 8e-05,
1079
+ 8e-05,
1080
+ 8e-05,
1081
+ 8e-05,
1082
+ 8e-05,
1083
+ 8e-05,
1084
+ 7.987186923279727e-05,
1085
+ 7.868547324017929e-05,
1086
+ 7.74990772475613e-05,
1087
+ 7.631268125494332e-05,
1088
+ 7.512628526232533e-05,
1089
+ 7.393988926970736e-05,
1090
+ 7.178064856314262e-05,
1091
+ 7.059425257052466e-05,
1092
+ 6.940785657790668e-05,
1093
+ 6.822146058528868e-05,
1094
+ 6.703506459267071e-05,
1095
+ 6.584866860005273e-05,
1096
+ 6.466227260743475e-05,
1097
+ 6.250303190087003e-05,
1098
+ 6.131663590825203e-05,
1099
+ 6.0130239915634074e-05,
1100
+ 5.894384392301608e-05,
1101
+ 5.77574479303981e-05,
1102
+ 5.657105193778013e-05,
1103
+ 5.538465594516214e-05,
1104
+ 5.3225415238597426e-05,
1105
+ 5.203901924597944e-05,
1106
+ 5.0852623253361464e-05,
1107
+ 4.966622726074349e-05,
1108
+ 4.8479831268125495e-05,
1109
+ 4.729343527550752e-05,
1110
+ 4.610703928288953e-05,
1111
+ 4.3947798576324816e-05,
1112
+ 4.276140258370682e-05,
1113
+ 4.157500659108885e-05,
1114
+ 4.038861059847088e-05,
1115
+ 3.9202214605852886e-05,
1116
+ 3.801581861323492e-05,
1117
+ 3.6829422620616924e-05,
1118
+ 3.46701819140522e-05,
1119
+ 3.348378592143423e-05,
1120
+ 3.229738992881624e-05,
1121
+ 3.111099393619827e-05,
1122
+ 2.9924597943580276e-05,
1123
+ 2.8738201950962302e-05,
1124
+ 2.7551805958344328e-05,
1125
+ 2.5392565251779594e-05,
1126
+ 2.4206169259161623e-05,
1127
+ 2.3019773266543632e-05,
1128
+ 2.1833377273925658e-05,
1129
+ 2.0646981281307667e-05,
1130
+ 1.9460585288689693e-05,
1131
+ 1.8274189296071722e-05,
1132
+ 1.6114948589506984e-05,
1133
+ 1.4928552596889012e-05,
1134
+ 1.374215660427102e-05,
1135
+ 1.2555760611653047e-05,
1136
+ 1.1369364619035074e-05,
1137
+ 1.0182968626417082e-05,
1138
+ 8.99657263379911e-06
1139
  ],
1140
  "emb_lr": [],
1141
  "eval_step": [
 
1148
  2696,
1149
  3087,
1150
  3478,
1151
+ 3869,
1152
+ 4260,
1153
+ 4651,
1154
+ 5042,
1155
+ 5433,
1156
+ 5824,
1157
+ 6215,
1158
+ 6606,
1159
+ 6997,
1160
+ 7388,
1161
+ 7779
1162
  ],
1163
  "eval_accuracy": [
1164
  0.01,
 
1170
  0.0,
1171
  0.0,
1172
  0.0,
1173
+ 0.0,
1174
+ 0.0,
1175
+ 0.0,
1176
+ 0.0,
1177
+ 0.0,
1178
+ 0.0,
1179
+ 0.0,
1180
+ 0.0,
1181
+ 0.0,
1182
+ 0.0,
1183
  0.0
1184
  ]
1185
  },
1186
+ "final_accuracy": 0.985,
1187
  "sft_eval": {
1188
  "config": {
1189
  "ops": "add_sub",
 
1194
  },
1195
  "splits": {
1196
  "add_S0": {
1197
+ "full_accuracy": 1.0,
1198
  "n_examples": 100,
1199
  "per_subtask": {
1200
  "SA": {
 
1202
  "count": 605
1203
  },
1204
  "SS": {
1205
+ "accuracy": 1.0,
1206
  "count": 95
1207
  }
1208
  }
1209
  },
1210
  "add_S1": {
1211
+ "full_accuracy": 1.0,
1212
  "n_examples": 100,
1213
  "per_subtask": {
1214
  "SA": {
1215
+ "accuracy": 1.0,
1216
  "count": 204
1217
  },
1218
  "SC": {
 
1230
  }
1231
  },
1232
  "add_S2": {
1233
+ "full_accuracy": 1.0,
1234
  "n_examples": 100,
1235
  "per_subtask": {
1236
  "SA": {
 
1238
  "count": 163
1239
  },
1240
  "SC": {
1241
+ "accuracy": 1.0,
1242
  "count": 130
1243
  },
1244
  "SS": {
1245
+ "accuracy": 1.0,
1246
  "count": 87
1247
  },
1248
  "UC": {
1249
+ "accuracy": 1.0,
1250
  "count": 203
1251
  },
1252
  "US": {
1253
+ "accuracy": 1.0,
1254
  "count": 117
1255
  }
1256
  }
1257
  },
1258
  "add_S3": {
1259
+ "full_accuracy": 0.98,
1260
  "n_examples": 100,
1261
  "per_subtask": {
1262
  "SA": {
1263
+ "accuracy": 1.0,
1264
  "count": 121
1265
  },
1266
  "SC": {
1267
+ "accuracy": 1.0,
1268
  "count": 121
1269
  },
1270
  "SS": {
 
1272
  "count": 49
1273
  },
1274
  "UC": {
1275
+ "accuracy": 0.989247311827957,
1276
  "count": 186
1277
  },
1278
  "US": {
1279
+ "accuracy": 1.0,
1280
  "count": 223
1281
  }
1282
  }
1283
  },
1284
  "add_S4": {
1285
+ "full_accuracy": 0.76,
1286
  "n_examples": 100,
1287
  "per_subtask": {
1288
  "SA": {
 
1290
  "count": 104
1291
  },
1292
  "SC": {
1293
+ "accuracy": 1.0,
1294
  "count": 106
1295
  },
1296
  "SS": {
 
1298
  "count": 23
1299
  },
1300
  "UC": {
1301
+ "accuracy": 0.85,
1302
  "count": 160
1303
  },
1304
  "US": {
1305
+ "accuracy": 0.9837133550488599,
1306
  "count": 307
1307
  }
1308
  }
1309
  },
1310
  "add_S5": {
1311
+ "full_accuracy": 0.45,
1312
  "n_examples": 100,
1313
  "per_subtask": {
1314
  "SA": {
 
1320
  "count": 100
1321
  },
1322
  "UC": {
1323
+ "accuracy": 0.53,
1324
  "count": 100
1325
  },
1326
  "US": {
1327
+ "accuracy": 0.8675,
1328
  "count": 400
1329
  }
1330
  }
1331
  },
1332
  "add_S6": {
1333
+ "full_accuracy": 0.49,
1334
  "n_examples": 100,
1335
  "per_subtask": {
1336
  "SC": {
 
1338
  "count": 100
1339
  },
1340
  "UC": {
1341
+ "accuracy": 0.65,
1342
  "count": 100
1343
  },
1344
  "US": {
1345
+ "accuracy": 0.824,
1346
  "count": 500
1347
  }
1348
  }
1349
  },
1350
  "add_random": {
1351
+ "full_accuracy": 0.99,
1352
  "n_examples": 200,
1353
  "per_subtask": {
1354
  "SA": {
1355
+ "accuracy": 1.0,
1356
  "count": 447
1357
  },
1358
  "SC": {
1359
+ "accuracy": 1.0,
1360
  "count": 320
1361
  },
1362
  "SS": {
 
1364
  "count": 56
1365
  },
1366
  "UC": {
1367
+ "accuracy": 0.996219281663516,
1368
  "count": 529
1369
  },
1370
  "US": {
1371
+ "accuracy": 1.0,
1372
  "count": 48
1373
  }
1374
  }
1375
  },
1376
  "add_C3": {
1377
+ "full_accuracy": 0.97,
1378
  "n_examples": 100,
1379
  "per_subtask": {
1380
  "SA": {
 
1386
  "count": 100
1387
  },
1388
  "UC": {
1389
+ "accuracy": 0.9844559585492227,
1390
  "count": 193
1391
  },
1392
  "US": {
1393
+ "accuracy": 1.0,
1394
  "count": 107
1395
  }
1396
  }
1397
  },
1398
  "add_C4": {
1399
+ "full_accuracy": 0.92,
1400
  "n_examples": 100,
1401
  "per_subtask": {
1402
  "SA": {
 
1408
  "count": 100
1409
  },
1410
  "UC": {
1411
+ "accuracy": 0.97265625,
1412
  "count": 256
1413
  },
1414
  "US": {
1415
+ "accuracy": 0.9861111111111112,
1416
  "count": 144
1417
  }
1418
  }
1419
  },
1420
  "add_C5": {
1421
+ "full_accuracy": 0.9,
1422
  "n_examples": 100,
1423
  "per_subtask": {
1424
  "SA": {
 
1430
  "count": 100
1431
  },
1432
  "UC": {
1433
+ "accuracy": 0.9673202614379085,
1434
  "count": 306
1435
  },
1436
  "US": {
1437
+ "accuracy": 0.979381443298969,
1438
  "count": 194
1439
  }
1440
  }
1441
  },
1442
  "add_C6": {
1443
+ "full_accuracy": 0.97,
1444
  "n_examples": 100,
1445
  "per_subtask": {
1446
  "SC": {
 
1448
  "count": 100
1449
  },
1450
  "UC": {
1451
+ "accuracy": 0.9918032786885246,
1452
  "count": 366
1453
  },
1454
  "US": {
1455
+ "accuracy": 1.0,
1456
  "count": 234
1457
  }
1458
  }
1459
  },
1460
  "sub_M0": {
1461
+ "full_accuracy": 0.99,
1462
  "n_examples": 100,
1463
  "per_subtask": {
1464
  "MD": {
1465
+ "accuracy": 0.9983361064891847,
1466
  "count": 601
1467
  },
1468
  "ME": {
 
1472
  }
1473
  },
1474
  "sub_M1": {
1475
+ "full_accuracy": 1.0,
1476
  "n_examples": 100,
1477
  "per_subtask": {
1478
  "MD": {
1479
+ "accuracy": 1.0,
1480
  "count": 279
1481
  },
1482
  "MB": {
1483
+ "accuracy": 1.0,
1484
  "count": 145
1485
  },
1486
  "ME": {
1487
+ "accuracy": 1.0,
1488
  "count": 24
1489
  },
1490
  "UB": {
1491
+ "accuracy": 1.0,
1492
  "count": 252
1493
  }
1494
  }
1495
  },
1496
  "sub_M2": {
1497
+ "full_accuracy": 1.0,
1498
  "n_examples": 100,
1499
  "per_subtask": {
1500
  "MD": {
 
1502
  "count": 213
1503
  },
1504
  "MB": {
1505
+ "accuracy": 1.0,
1506
  "count": 113
1507
  },
1508
  "ME": {
 
1510
  "count": 85
1511
  },
1512
  "UB": {
1513
+ "accuracy": 1.0,
1514
  "count": 181
1515
  },
1516
  "UD": {
 
1520
  }
1521
  },
1522
  "sub_M3": {
1523
+ "full_accuracy": 0.99,
1524
  "n_examples": 100,
1525
  "per_subtask": {
1526
  "MD": {
 
1536
  "count": 56
1537
  },
1538
  "UB": {
1539
+ "accuracy": 0.9932885906040269,
1540
  "count": 149
1541
  },
1542
  "UD": {
1543
+ "accuracy": 1.0,
1544
  "count": 213
1545
  }
1546
  }
1547
  },
1548
  "sub_M4": {
1549
+ "full_accuracy": 0.85,
1550
  "n_examples": 100,
1551
  "per_subtask": {
1552
  "MD": {
 
1558
  "count": 100
1559
  },
1560
  "UB": {
1561
+ "accuracy": 0.85,
1562
  "count": 100
1563
  },
1564
  "UD": {
1565
+ "accuracy": 1.0,
1566
  "count": 300
1567
  }
1568
  }
1569
  },
1570
  "sub_M5": {
1571
+ "full_accuracy": 0.73,
1572
  "n_examples": 100,
1573
  "per_subtask": {
1574
  "MD": {
 
1580
  "count": 100
1581
  },
1582
  "UB": {
1583
+ "accuracy": 0.95,
1584
  "count": 100
1585
  },
1586
  "UD": {
1587
+ "accuracy": 0.945,
1588
  "count": 400
1589
  }
1590
  }
1591
  },
1592
  "sub_random": {
1593
+ "full_accuracy": 0.985,
1594
  "n_examples": 200,
1595
  "per_subtask": {
1596
  "MD": {
 
1606
  "count": 53
1607
  },
1608
  "UB": {
1609
+ "accuracy": 1.0,
1610
  "count": 439
1611
  },
1612
  "UD": {
 
1616
  }
1617
  },
1618
  "sub_B3": {
1619
+ "full_accuracy": 0.97,
1620
  "n_examples": 100,
1621
  "per_subtask": {
1622
  "MD": {
1623
+ "accuracy": 0.9933333333333333,
1624
  "count": 300
1625
  },
1626
  "MB": {
 
1628
  "count": 100
1629
  },
1630
  "UB": {
1631
+ "accuracy": 0.9949238578680203,
1632
  "count": 197
1633
  },
1634
  "UD": {
1635
+ "accuracy": 1.0,
1636
  "count": 103
1637
  }
1638
  }
1639
  },
1640
  "sub_B4": {
1641
+ "full_accuracy": 0.98,
1642
  "n_examples": 100,
1643
  "per_subtask": {
1644
  "MD": {
 
1650
  "count": 100
1651
  },
1652
  "UB": {
1653
+ "accuracy": 0.9959514170040485,
1654
  "count": 247
1655
  },
1656
  "UD": {
1657
+ "accuracy": 0.9934640522875817,
1658
  "count": 153
1659
  }
1660
  }
1661
  },
1662
  "sub_B5": {
1663
+ "full_accuracy": 1.0,
1664
  "n_examples": 100,
1665
  "per_subtask": {
1666
  "MD": {
 
1672
  "count": 100
1673
  },
1674
  "UB": {
1675
+ "accuracy": 1.0,
1676
  "count": 298
1677
  },
1678
  "UD": {
1679
+ "accuracy": 1.0,
1680
  "count": 202
1681
  }
1682
  }
1683
  }
1684
  },
1685
  "summary": {
1686
+ "overall_accuracy": 0.9125,
1687
  "total_examples": 2400,
1688
  "n_splits": 22
1689
  }
 
1698
  },
1699
  "splits": {
1700
  "add_S0": {
1701
+ "full_accuracy": 1.0,
1702
  "n_examples": 100,
1703
  "per_subtask": {
1704
  "SA": {
1705
+ "accuracy": 1.0,
1706
  "count": 605
1707
  },
1708
  "SS": {
 
1734
  }
1735
  },
1736
  "add_S2": {
1737
+ "full_accuracy": 1.0,
1738
  "n_examples": 100,
1739
  "per_subtask": {
1740
  "SA": {
 
1750
  "count": 87
1751
  },
1752
  "UC": {
1753
+ "accuracy": 1.0,
1754
  "count": 203
1755
  },
1756
  "US": {
 
1760
  }
1761
  },
1762
  "add_S3": {
1763
+ "full_accuracy": 1.0,
1764
  "n_examples": 100,
1765
  "per_subtask": {
1766
  "SA": {
 
1776
  "count": 49
1777
  },
1778
  "UC": {
1779
+ "accuracy": 1.0,
1780
  "count": 186
1781
  },
1782
  "US": {
 
1786
  }
1787
  },
1788
  "add_S4": {
1789
+ "full_accuracy": 1.0,
1790
  "n_examples": 100,
1791
  "per_subtask": {
1792
  "SA": {
 
1802
  "count": 23
1803
  },
1804
  "UC": {
1805
+ "accuracy": 1.0,
1806
  "count": 160
1807
  },
1808
  "US": {
1809
+ "accuracy": 1.0,
1810
  "count": 307
1811
  }
1812
  }
1813
  },
1814
  "add_S5": {
1815
+ "full_accuracy": 0.96,
1816
  "n_examples": 100,
1817
  "per_subtask": {
1818
  "SA": {
 
1824
  "count": 100
1825
  },
1826
  "UC": {
1827
+ "accuracy": 0.96,
1828
  "count": 100
1829
  },
1830
  "US": {
1831
+ "accuracy": 1.0,
1832
  "count": 400
1833
  }
1834
  }
1835
  },
1836
  "add_S6": {
1837
+ "full_accuracy": 0.91,
1838
  "n_examples": 100,
1839
  "per_subtask": {
1840
  "SC": {
 
1842
  "count": 100
1843
  },
1844
  "UC": {
1845
+ "accuracy": 0.91,
1846
  "count": 100
1847
  },
1848
  "US": {
1849
+ "accuracy": 0.984,
1850
  "count": 500
1851
  }
1852
  }
1853
  },
1854
  "add_random": {
1855
+ "full_accuracy": 1.0,
1856
  "n_examples": 200,
1857
  "per_subtask": {
1858
  "SA": {
1859
+ "accuracy": 1.0,
1860
  "count": 447
1861
  },
1862
  "SC": {
1863
+ "accuracy": 1.0,
1864
  "count": 320
1865
  },
1866
  "SS": {
 
1878
  }
1879
  },
1880
  "add_C3": {
1881
+ "full_accuracy": 1.0,
1882
  "n_examples": 100,
1883
  "per_subtask": {
1884
  "SA": {
 
1890
  "count": 100
1891
  },
1892
  "UC": {
1893
+ "accuracy": 1.0,
1894
  "count": 193
1895
  },
1896
  "US": {
 
1900
  }
1901
  },
1902
  "add_C4": {
1903
+ "full_accuracy": 1.0,
1904
  "n_examples": 100,
1905
  "per_subtask": {
1906
  "SA": {
 
1912
  "count": 100
1913
  },
1914
  "UC": {
1915
+ "accuracy": 1.0,
1916
  "count": 256
1917
  },
1918
  "US": {
1919
+ "accuracy": 1.0,
1920
  "count": 144
1921
  }
1922
  }
1923
  },
1924
  "add_C5": {
1925
+ "full_accuracy": 1.0,
1926
  "n_examples": 100,
1927
  "per_subtask": {
1928
  "SA": {
 
1934
  "count": 100
1935
  },
1936
  "UC": {
1937
+ "accuracy": 1.0,
1938
  "count": 306
1939
  },
1940
  "US": {
1941
+ "accuracy": 1.0,
1942
  "count": 194
1943
  }
1944
  }
1945
  },
1946
  "add_C6": {
1947
+ "full_accuracy": 0.99,
1948
  "n_examples": 100,
1949
  "per_subtask": {
1950
  "SC": {
 
1952
  "count": 100
1953
  },
1954
  "UC": {
1955
+ "accuracy": 0.9972677595628415,
1956
  "count": 366
1957
  },
1958
  "US": {
1959
+ "accuracy": 1.0,
1960
  "count": 234
1961
  }
1962
  }
 
2024
  }
2025
  },
2026
  "sub_M3": {
2027
+ "full_accuracy": 1.0,
2028
  "n_examples": 100,
2029
  "per_subtask": {
2030
  "MD": {
 
2040
  "count": 56
2041
  },
2042
  "UB": {
2043
+ "accuracy": 1.0,
2044
  "count": 149
2045
  },
2046
  "UD": {
2047
+ "accuracy": 1.0,
2048
  "count": 213
2049
  }
2050
  }
2051
  },
2052
  "sub_M4": {
2053
+ "full_accuracy": 1.0,
2054
  "n_examples": 100,
2055
  "per_subtask": {
2056
  "MD": {
 
2062
  "count": 100
2063
  },
2064
  "UB": {
2065
+ "accuracy": 1.0,
2066
  "count": 100
2067
  },
2068
  "UD": {
2069
+ "accuracy": 1.0,
2070
  "count": 300
2071
  }
2072
  }
2073
  },
2074
  "sub_M5": {
2075
+ "full_accuracy": 0.8,
2076
  "n_examples": 100,
2077
  "per_subtask": {
2078
  "MD": {
 
2084
  "count": 100
2085
  },
2086
  "UB": {
2087
+ "accuracy": 0.8,
2088
  "count": 100
2089
  },
2090
  "UD": {
2091
+ "accuracy": 0.9925,
2092
  "count": 400
2093
  }
2094
  }
2095
  },
2096
  "sub_random": {
2097
+ "full_accuracy": 0.995,
2098
  "n_examples": 200,
2099
  "per_subtask": {
2100
  "MD": {
2101
+ "accuracy": 0.9983333333333333,
2102
  "count": 600
2103
  },
2104
  "MB": {
 
2120
  }
2121
  },
2122
  "sub_B3": {
2123
+ "full_accuracy": 1.0,
2124
  "n_examples": 100,
2125
  "per_subtask": {
2126
  "MD": {
 
2132
  "count": 100
2133
  },
2134
  "UB": {
2135
+ "accuracy": 1.0,
2136
  "count": 197
2137
  },
2138
  "UD": {
2139
+ "accuracy": 1.0,
2140
  "count": 103
2141
  }
2142
  }
2143
  },
2144
  "sub_B4": {
2145
+ "full_accuracy": 1.0,
2146
  "n_examples": 100,
2147
  "per_subtask": {
2148
  "MD": {
 
2154
  "count": 100
2155
  },
2156
  "UB": {
2157
+ "accuracy": 1.0,
2158
  "count": 247
2159
  },
2160
  "UD": {
2161
+ "accuracy": 1.0,
2162
  "count": 153
2163
  }
2164
  }
2165
  },
2166
  "sub_B5": {
2167
+ "full_accuracy": 0.99,
2168
  "n_examples": 100,
2169
  "per_subtask": {
2170
  "MD": {
 
2176
  "count": 100
2177
  },
2178
  "UB": {
2179
+ "accuracy": 0.9966442953020134,
2180
  "count": 298
2181
  },
2182
  "UD": {
2183
+ "accuracy": 1.0,
2184
  "count": 202
2185
  }
2186
  }
2187
  }
2188
  },
2189
  "summary": {
2190
+ "overall_accuracy": 0.985,
2191
  "total_examples": 2400,
2192
  "n_splits": 22
2193
  }
2194
  },
2195
+ "sorl_overall_accuracy": 0.985,
2196
+ "sft_overall_accuracy": 0.9125
2197
  }
add_sub_sorl_v1_abs10_K1_25K/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:cf047dee834e21f59a68ae2338d8d11dc22cd4fd412eacace785b4753ce6d636
3
  size 650303660
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:17d304a942f99c334b066a90cb3feae8b0860d4a67f8cf4d6fc325dad7517635
3
  size 650303660
add_sub_sorl_v1_abs10_K1_25K/train_config.json CHANGED
@@ -20,7 +20,7 @@
20
  "lr": 8e-05,
21
  "emb_lr_mult": 1.0,
22
  "weight_decay": 0.01,
23
- "warmup_steps": 117,
24
  "cooldown_frac": 0.4,
25
  "max_grad_norm": 1.0,
26
  "vq_abs_pretrain_steps": 0,
@@ -30,7 +30,7 @@
30
  "vq_abs_pretrain_target_vectors": 20000,
31
  "batch_size": 64,
32
  "gradient_accumulation_steps": 1,
33
- "num_epochs": 10,
34
  "emb_warmup_steps": 0,
35
  "log_every": 50,
36
  "eval_every": 390,
@@ -69,16 +69,16 @@
69
  "no_wandb": false,
70
  "n_params": 162499262,
71
  "run_name": "add_sub_sorl_v1_abs10_K1_25K",
72
- "git_commit": "8d5ee5420119746ef4e2c87570eb250c9718f643",
73
- "timestamp": "2026-04-12T21:07:34.657454+00:00",
74
  "tokenizer": "Qwen/Qwen3-0.6B",
75
  "dataset_repo": "thoughtworks/arithmetic-sorl-data",
76
  "dataset_config": "add_sub_6digit",
77
  "model_repo": "thoughtworks/arithmetic-sorl",
78
  "trainer_version": "v1",
79
- "wandb_run_id": "ihgpsgwk",
80
- "wandb_url": "https://wandb.ai/nlp_and_interpretability/sorl-arithmetic/runs/ihgpsgwk",
81
- "final_accuracy": 0.8683333333333333,
82
- "sft_accuracy": 0.66375,
83
  "eval_method": "ArithmeticEvaluator"
84
  }
 
20
  "lr": 8e-05,
21
  "emb_lr_mult": 1.0,
22
  "weight_decay": 0.01,
23
+ "warmup_steps": 234,
24
  "cooldown_frac": 0.4,
25
  "max_grad_norm": 1.0,
26
  "vq_abs_pretrain_steps": 0,
 
30
  "vq_abs_pretrain_target_vectors": 20000,
31
  "batch_size": 64,
32
  "gradient_accumulation_steps": 1,
33
+ "num_epochs": 20,
34
  "emb_warmup_steps": 0,
35
  "log_every": 50,
36
  "eval_every": 390,
 
69
  "no_wandb": false,
70
  "n_params": 162499262,
71
  "run_name": "add_sub_sorl_v1_abs10_K1_25K",
72
+ "git_commit": "57deaa28d9c21e39ddac5ef448d6e1be992fba91",
73
+ "timestamp": "2026-04-13T11:27:47.069069+00:00",
74
  "tokenizer": "Qwen/Qwen3-0.6B",
75
  "dataset_repo": "thoughtworks/arithmetic-sorl-data",
76
  "dataset_config": "add_sub_6digit",
77
  "model_repo": "thoughtworks/arithmetic-sorl",
78
  "trainer_version": "v1",
79
+ "wandb_run_id": "swo4evkc",
80
+ "wandb_url": "https://wandb.ai/nlp_and_interpretability/sorl-arithmetic/runs/swo4evkc",
81
+ "final_accuracy": 0.985,
82
+ "sft_accuracy": 0.9125,
83
  "eval_method": "ArithmeticEvaluator"
84
  }