amirali1985 commited on
Commit
d7d9f7c
·
verified ·
1 Parent(s): 9cdedc9

Upload add_sub_baseline_100K_1L2H256d

Browse files
add_sub_baseline_100K_1L2H256d/metrics.json CHANGED
@@ -628,1258 +628,1258 @@
628
  31250
629
  ],
630
  "loss": [
631
- 11.928034782409668,
632
- 11.653544425964355,
633
- 11.307957649230957,
634
- 10.721878051757812,
635
- 10.181405067443848,
636
- 9.667264938354492,
637
- 9.305678367614746,
638
- 9.201789855957031,
639
- 8.780132293701172,
640
- 8.337949752807617,
641
- 8.104902267456055,
642
- 7.672562599182129,
643
- 7.409525394439697,
644
- 6.95243501663208,
645
- 6.568300724029541,
646
- 6.088183879852295,
647
- 5.540016174316406,
648
- 5.10260534286499,
649
- 4.484571933746338,
650
- 4.0001749992370605,
651
- 3.444486141204834,
652
- 3.0357420444488525,
653
- 2.7352375984191895,
654
- 2.3701672554016113,
655
- 2.2287418842315674,
656
- 2.0617270469665527,
657
- 2.0463647842407227,
658
- 1.8895787000656128,
659
- 1.8856637477874756,
660
- 1.9227359294891357,
661
- 1.8609199523925781,
662
- 1.9265002012252808,
663
- 1.9278079271316528,
664
- 1.8539155721664429,
665
- 1.8715951442718506,
666
- 1.7844884395599365,
667
- 1.8312513828277588,
668
- 1.7931902408599854,
669
- 1.810530424118042,
670
- 1.8436228036880493,
671
- 1.8527024984359741,
672
- 1.8303594589233398,
673
- 1.7831075191497803,
674
- 1.7217122316360474,
675
- 1.7545881271362305,
676
- 1.7245683670043945,
677
- 1.6838029623031616,
678
- 1.7725518941879272,
679
- 1.7447564601898193,
680
- 1.6834079027175903,
681
- 1.7119642496109009,
682
- 1.7137107849121094,
683
- 1.7549256086349487,
684
- 1.7131061553955078,
685
- 1.7283432483673096,
686
- 1.7024879455566406,
687
- 1.6835523843765259,
688
- 1.724271535873413,
689
- 1.6476503610610962,
690
- 1.6461162567138672,
691
- 1.6732215881347656,
692
- 1.6631882190704346,
693
- 1.6530768871307373,
694
- 1.640184998512268,
695
- 1.5616495609283447,
696
- 1.5732653141021729,
697
- 1.5028226375579834,
698
- 1.469290852546692,
699
- 1.4607034921646118,
700
- 1.4396305084228516,
701
- 1.33017897605896,
702
- 1.173120141029358,
703
- 1.0654069185256958,
704
- 1.1612730026245117,
705
- 1.1505136489868164,
706
- 1.0570982694625854,
707
- 0.9696107506752014,
708
- 0.9057245254516602,
709
- 0.8982640504837036,
710
- 0.8462095856666565,
711
- 0.7811938524246216,
712
- 0.728919506072998,
713
- 0.7578886151313782,
714
- 0.7812106013298035,
715
- 0.7021167874336243,
716
- 0.7130362391471863,
717
- 0.6820688247680664,
718
- 0.6510214805603027,
719
- 0.6453285217285156,
720
- 0.6750946640968323,
721
- 0.6662521362304688,
722
- 0.589103102684021,
723
- 0.635167121887207,
724
- 0.6322607398033142,
725
- 0.6354484558105469,
726
- 0.5717305541038513,
727
- 0.5770869851112366,
728
- 0.5638580322265625,
729
- 0.6003702282905579,
730
- 0.606711208820343,
731
- 0.5978370308876038,
732
- 0.5625749826431274,
733
- 0.5271865725517273,
734
- 0.5920223593711853,
735
- 0.6064496636390686,
736
- 0.5727962255477905,
737
- 0.5373821258544922,
738
- 0.547116219997406,
739
- 0.5527753233909607,
740
- 0.5407260060310364,
741
- 0.5330578088760376,
742
- 0.5460835695266724,
743
- 0.5407729744911194,
744
- 0.5695903301239014,
745
- 0.546602725982666,
746
- 0.5379611253738403,
747
- 0.49000412225723267,
748
- 0.5377251505851746,
749
- 0.5077806711196899,
750
- 0.5233360528945923,
751
- 0.5038724541664124,
752
- 0.5082470774650574,
753
- 0.5090785622596741,
754
- 0.4877682328224182,
755
- 0.46342095732688904,
756
- 0.4934587776660919,
757
- 0.47618719935417175,
758
- 0.5213378071784973,
759
- 0.48468828201293945,
760
- 0.4764997959136963,
761
- 0.490345299243927,
762
- 0.48699042201042175,
763
- 0.44104668498039246,
764
- 0.474947452545166,
765
- 0.46392837166786194,
766
- 0.4587855041027069,
767
- 0.4935186207294464,
768
- 0.46145424246788025,
769
- 0.4339943826198578,
770
- 0.4798485338687897,
771
- 0.4835148751735687,
772
- 0.45113644003868103,
773
- 0.4499821066856384,
774
- 0.44402727484703064,
775
- 0.4216662049293518,
776
- 0.4317099452018738,
777
- 0.4295271337032318,
778
- 0.43690910935401917,
779
- 0.4310178756713867,
780
- 0.42988887429237366,
781
- 0.39782506227493286,
782
- 0.40841910243034363,
783
- 0.4422263205051422,
784
- 0.4496307075023651,
785
- 0.4485880434513092,
786
- 0.43126311898231506,
787
- 0.41150495409965515,
788
- 0.4591612219810486,
789
- 0.45604208111763,
790
- 0.46709761023521423,
791
- 0.4348989427089691,
792
- 0.3834414482116699,
793
- 0.38175979256629944,
794
- 0.3879620134830475,
795
- 0.40807420015335083,
796
- 0.4096015989780426,
797
- 0.4135327637195587,
798
- 0.3968922793865204,
799
- 0.43453529477119446,
800
- 0.36617913842201233,
801
- 0.41760990023612976,
802
- 0.3654819130897522,
803
- 0.43267178535461426,
804
- 0.3825819492340088,
805
- 0.40239834785461426,
806
- 0.4138084053993225,
807
- 0.39564505219459534,
808
- 0.36791282892227173,
809
- 0.43040919303894043,
810
- 0.4024931490421295,
811
- 0.3815248906612396,
812
- 0.40737414360046387,
813
- 0.42400693893432617,
814
- 0.38643333315849304,
815
- 0.3655295670032501,
816
- 0.3794443905353546,
817
- 0.3761146664619446,
818
- 0.402232825756073,
819
- 0.39684897661209106,
820
- 0.41486597061157227,
821
- 0.3713434338569641,
822
- 0.4122335612773895,
823
- 0.3593781888484955,
824
- 0.36669155955314636,
825
- 0.38098907470703125,
826
- 0.3585493564605713,
827
- 0.3658697009086609,
828
- 0.3952808976173401,
829
- 0.3677598536014557,
830
- 0.40938600897789,
831
- 0.3865565359592438,
832
- 0.3737422823905945,
833
- 0.36394765973091125,
834
- 0.33259841799736023,
835
- 0.3456837832927704,
836
- 0.33143261075019836,
837
- 0.35574594140052795,
838
- 0.3818463683128357,
839
- 0.32026952505111694,
840
- 0.35572943091392517,
841
- 0.3279140889644623,
842
- 0.38622188568115234,
843
- 0.32534393668174744,
844
- 0.35246992111206055,
845
- 0.34478285908699036,
846
- 0.3526977002620697,
847
- 0.3618982434272766,
848
- 0.3712795674800873,
849
- 0.3642587661743164,
850
- 0.3191487491130829,
851
- 0.36171674728393555,
852
- 0.34270623326301575,
853
- 0.3760503828525543,
854
- 0.34943634271621704,
855
- 0.3346637785434723,
856
- 0.32916831970214844,
857
- 0.3213987350463867,
858
- 0.34669747948646545,
859
- 0.34980323910713196,
860
- 0.3436082899570465,
861
- 0.3271951675415039,
862
- 0.31385931372642517,
863
- 0.3484307825565338,
864
- 0.3084122836589813,
865
- 0.3468055725097656,
866
- 0.3438337743282318,
867
- 0.3528638482093811,
868
- 0.31498146057128906,
869
- 0.3160567879676819,
870
- 0.34172186255455017,
871
- 0.3417809009552002,
872
- 0.3534678816795349,
873
- 0.35489383339881897,
874
- 0.34530016779899597,
875
- 0.2865831255912781,
876
- 0.3117128908634186,
877
- 0.34365028142929077,
878
- 0.33491435647010803,
879
- 0.3689354360103607,
880
- 0.3382095396518707,
881
- 0.3239511549472809,
882
- 0.3418653905391693,
883
- 0.3021024167537689,
884
- 0.2963162362575531,
885
- 0.32997259497642517,
886
- 0.33671286702156067,
887
- 0.33611273765563965,
888
- 0.3459137976169586,
889
- 0.33286041021347046,
890
- 0.2873530983924866,
891
- 0.31002670526504517,
892
- 0.30808526277542114,
893
- 0.287386417388916,
894
- 0.30139049887657166,
895
- 0.29464051127433777,
896
- 0.3082091510295868,
897
- 0.3462357223033905,
898
- 0.34610894322395325,
899
- 0.3279619514942169,
900
- 0.3180283010005951,
901
- 0.2845425009727478,
902
- 0.27627280354499817,
903
- 0.29058805108070374,
904
- 0.25387290120124817,
905
- 0.29174312949180603,
906
- 0.3285600244998932,
907
- 0.31100231409072876,
908
- 0.3193957507610321,
909
- 0.3087748885154724,
910
- 0.3061353862285614,
911
- 0.32200178503990173,
912
- 0.2894422709941864,
913
- 0.2870963513851166,
914
- 0.31858688592910767,
915
- 0.2746082842350006,
916
- 0.3215934932231903,
917
- 0.3257999122142792,
918
- 0.31921395659446716,
919
- 0.2667975425720215,
920
- 0.3186761736869812,
921
- 0.2847942113876343,
922
- 0.28523775935173035,
923
- 0.32166609168052673,
924
- 0.3076777458190918,
925
- 0.3031381666660309,
926
- 0.29915809631347656,
927
- 0.32860395312309265,
928
- 0.3052523136138916,
929
- 0.31743431091308594,
930
- 0.27312588691711426,
931
- 0.26801934838294983,
932
- 0.26507896184921265,
933
- 0.294921338558197,
934
- 0.3115091919898987,
935
- 0.26985788345336914,
936
- 0.293649286031723,
937
- 0.2707940638065338,
938
- 0.29856735467910767,
939
- 0.2623736262321472,
940
- 0.3006415367126465,
941
- 0.31037160754203796,
942
- 0.30440378189086914,
943
- 0.3023694157600403,
944
- 0.2869836390018463,
945
- 0.24748651683330536,
946
- 0.28105297684669495,
947
- 0.2706967294216156,
948
- 0.3016051948070526,
949
- 0.2754139006137848,
950
- 0.28681060671806335,
951
- 0.28182414174079895,
952
- 0.29593154788017273,
953
- 0.2871556580066681,
954
- 0.29613766074180603,
955
- 0.30644792318344116,
956
- 0.25587597489356995,
957
- 0.28945502638816833,
958
- 0.27285969257354736,
959
- 0.28049322962760925,
960
- 0.2884880602359772,
961
- 0.2916378378868103,
962
- 0.2754189372062683,
963
- 0.3063291907310486,
964
- 0.3094882071018219,
965
- 0.29223495721817017,
966
- 0.2699856460094452,
967
- 0.265081524848938,
968
- 0.3157065212726593,
969
- 0.2825828790664673,
970
- 0.2644478380680084,
971
- 0.2736837565898895,
972
- 0.2939130365848541,
973
- 0.27335861325263977,
974
- 0.27167394757270813,
975
- 0.2819110155105591,
976
- 0.289447158575058,
977
- 0.2571549415588379,
978
- 0.2593221962451935,
979
- 0.28376683592796326,
980
- 0.29852765798568726,
981
- 0.27658429741859436,
982
- 0.23924340307712555,
983
- 0.28641754388809204,
984
- 0.3209593594074249,
985
- 0.2669186592102051,
986
- 0.2522584795951843,
987
- 0.2631411552429199,
988
- 0.30004528164863586,
989
- 0.25664424896240234,
990
- 0.26312512159347534,
991
- 0.2851202189922333,
992
- 0.2921363413333893,
993
- 0.24688827991485596,
994
- 0.2517416179180145,
995
- 0.30236250162124634,
996
- 0.24213668704032898,
997
- 0.2524048686027527,
998
- 0.27556005120277405,
999
- 0.27097469568252563,
1000
- 0.29630640149116516,
1001
- 0.24852998554706573,
1002
- 0.29177477955818176,
1003
- 0.2854974567890167,
1004
- 0.2374695986509323,
1005
- 0.23592327535152435,
1006
- 0.3132108151912689,
1007
- 0.2631916403770447,
1008
- 0.2837892472743988,
1009
- 0.25606682896614075,
1010
- 0.27506956458091736,
1011
- 0.24950344860553741,
1012
- 0.2695748507976532,
1013
- 0.2633344829082489,
1014
- 0.2621081471443176,
1015
- 0.2588253915309906,
1016
- 0.26628708839416504,
1017
- 0.2546164393424988,
1018
- 0.2710006535053253,
1019
- 0.2939824163913727,
1020
- 0.23946473002433777,
1021
- 0.22712087631225586,
1022
- 0.2623310983181,
1023
- 0.2482786327600479,
1024
- 0.27607423067092896,
1025
- 0.25139644742012024,
1026
- 0.28528496623039246,
1027
- 0.24713945388793945,
1028
- 0.24412168562412262,
1029
- 0.25628599524497986,
1030
- 0.28033557534217834,
1031
- 0.26747894287109375,
1032
- 0.29976686835289,
1033
- 0.29141291975975037,
1034
- 0.2702557146549225,
1035
- 0.24962809681892395,
1036
- 0.2612250745296478,
1037
- 0.2625961899757385,
1038
- 0.2644568383693695,
1039
- 0.2803192436695099,
1040
- 0.23608256876468658,
1041
- 0.23363173007965088,
1042
- 0.2611221969127655,
1043
- 0.2525368332862854,
1044
- 0.2524564564228058,
1045
- 0.24044832587242126,
1046
- 0.2552793622016907,
1047
- 0.28079238533973694,
1048
- 0.2619774639606476,
1049
- 0.23810975253582,
1050
- 0.24849574267864227,
1051
- 0.25039538741111755,
1052
- 0.2613367438316345,
1053
- 0.2615041732788086,
1054
- 0.20929165184497833,
1055
- 0.23789460957050323,
1056
- 0.2508893311023712,
1057
- 0.23562470078468323,
1058
- 0.27942875027656555,
1059
- 0.282475084066391,
1060
- 0.26715087890625,
1061
- 0.25968289375305176,
1062
- 0.26287397742271423,
1063
- 0.25527840852737427,
1064
- 0.24361786246299744,
1065
- 0.24610944092273712,
1066
- 0.2573030889034271,
1067
- 0.28686219453811646,
1068
- 0.25092869997024536,
1069
- 0.26785707473754883,
1070
- 0.2560536861419678,
1071
- 0.2647089958190918,
1072
- 0.2567795217037201,
1073
- 0.2601569592952728,
1074
- 0.24640819430351257,
1075
- 0.2674255669116974,
1076
- 0.2340751439332962,
1077
- 0.22037099301815033,
1078
- 0.23103085160255432,
1079
- 0.22777055203914642,
1080
- 0.24295179545879364,
1081
- 0.23772236704826355,
1082
- 0.23818251490592957,
1083
- 0.24417351186275482,
1084
- 0.23382483422756195,
1085
- 0.2717232406139374,
1086
- 0.24655917286872864,
1087
- 0.24539153277873993,
1088
- 0.27003398537635803,
1089
- 0.24944210052490234,
1090
- 0.21866385638713837,
1091
- 0.24142920970916748,
1092
- 0.2841137945652008,
1093
- 0.27210989594459534,
1094
- 0.24623851478099823,
1095
- 0.2414173036813736,
1096
- 0.24330724775791168,
1097
- 0.2612263560295105,
1098
- 0.24263326823711395,
1099
- 0.23556897044181824,
1100
- 0.25661733746528625,
1101
- 0.23128819465637207,
1102
- 0.2515091300010681,
1103
- 0.23440249264240265,
1104
- 0.26295438408851624,
1105
- 0.24453571438789368,
1106
- 0.25462159514427185,
1107
- 0.23500898480415344,
1108
- 0.24620413780212402,
1109
- 0.23643255233764648,
1110
- 0.21674731373786926,
1111
- 0.23458817601203918,
1112
- 0.22282226383686066,
1113
- 0.2536497414112091,
1114
- 0.23018422722816467,
1115
- 0.21616289019584656,
1116
- 0.24523791670799255,
1117
- 0.2539035975933075,
1118
- 0.2417120784521103,
1119
- 0.2687821388244629,
1120
- 0.2591947615146637,
1121
- 0.23294435441493988,
1122
- 0.24325568974018097,
1123
- 0.2612447142601013,
1124
- 0.23927579820156097,
1125
- 0.23915018141269684,
1126
- 0.26768800616264343,
1127
- 0.24271290004253387,
1128
- 0.25881507992744446,
1129
- 0.23990657925605774,
1130
- 0.22135673463344574,
1131
- 0.20841874182224274,
1132
- 0.24352209270000458,
1133
- 0.2426811009645462,
1134
- 0.22436712682247162,
1135
- 0.24585120379924774,
1136
- 0.23543156683444977,
1137
- 0.24074861407279968,
1138
- 0.2525044083595276,
1139
- 0.24585461616516113,
1140
- 0.22054941952228546,
1141
- 0.23277811706066132,
1142
- 0.2542433738708496,
1143
- 0.21179088950157166,
1144
- 0.2887387275695801,
1145
- 0.24148423969745636,
1146
- 0.25414398312568665,
1147
- 0.25563937425613403,
1148
- 0.23944461345672607,
1149
- 0.2533447742462158,
1150
- 0.23563027381896973,
1151
- 0.26421570777893066,
1152
- 0.2254643440246582,
1153
- 0.24568693339824677,
1154
- 0.2360592931509018,
1155
- 0.2221137136220932,
1156
- 0.2282101809978485,
1157
- 0.24913156032562256,
1158
- 0.22136715054512024,
1159
- 0.23910680413246155,
1160
- 0.23922684788703918,
1161
- 0.22790290415287018,
1162
- 0.23584823310375214,
1163
- 0.2730078399181366,
1164
- 0.2663761079311371,
1165
- 0.248269185423851,
1166
- 0.25979501008987427,
1167
- 0.2336592972278595,
1168
- 0.2624785006046295,
1169
- 0.22696606814861298,
1170
- 0.21379384398460388,
1171
- 0.21497604250907898,
1172
- 0.2592029571533203,
1173
- 0.22469215095043182,
1174
- 0.25158485770225525,
1175
- 0.24422571063041687,
1176
- 0.25753626227378845,
1177
- 0.25007328391075134,
1178
- 0.2602040469646454,
1179
- 0.23332370817661285,
1180
- 0.23722335696220398,
1181
- 0.24797333776950836,
1182
- 0.22556526958942413,
1183
- 0.23421958088874817,
1184
- 0.24832820892333984,
1185
- 0.2311556339263916,
1186
- 0.23707018792629242,
1187
- 0.23783321678638458,
1188
- 0.20987555384635925,
1189
- 0.267695814371109,
1190
- 0.2629874646663666,
1191
- 0.21873648464679718,
1192
- 0.24494552612304688,
1193
- 0.26303473114967346,
1194
- 0.24542763829231262,
1195
- 0.23399154841899872,
1196
- 0.23643358051776886,
1197
- 0.21889302134513855,
1198
- 0.22333727777004242,
1199
- 0.20176778733730316,
1200
- 0.2578296363353729,
1201
- 0.26290076971054077,
1202
- 0.2378210574388504,
1203
- 0.24513255059719086,
1204
- 0.25607389211654663,
1205
- 0.24997523427009583,
1206
- 0.2558247745037079,
1207
- 0.25033673644065857,
1208
- 0.2546168267726898,
1209
- 0.23066361248493195,
1210
- 0.22089658677577972,
1211
- 0.2460862696170807,
1212
- 0.21666987240314484,
1213
- 0.22690625488758087,
1214
- 0.24581076204776764,
1215
- 0.2427884191274643,
1216
- 0.22972403466701508,
1217
- 0.25537118315696716,
1218
- 0.24844148755073547,
1219
- 0.2209678441286087,
1220
- 0.22279927134513855,
1221
- 0.23022934794425964,
1222
- 0.2374199479818344,
1223
- 0.2532387375831604,
1224
- 0.24079224467277527,
1225
- 0.2286282330751419,
1226
- 0.22575612366199493,
1227
- 0.22752560675144196,
1228
- 0.26504701375961304,
1229
- 0.23773540556430817,
1230
- 0.20410892367362976,
1231
- 0.25362756848335266,
1232
- 0.24984030425548553,
1233
- 0.2467348873615265,
1234
- 0.2259712964296341,
1235
- 0.23826143145561218,
1236
- 0.23323170840740204,
1237
- 0.23948879539966583,
1238
- 0.24911201000213623,
1239
- 0.23734129965305328,
1240
- 0.24040012061595917,
1241
- 0.26085561513900757,
1242
- 0.25282594561576843,
1243
- 0.24620071053504944,
1244
- 0.2408737689256668,
1245
- 0.22477446496486664,
1246
- 0.232040137052536,
1247
- 0.22941991686820984,
1248
- 0.2574242055416107,
1249
- 0.22758789360523224,
1250
- 0.2541619539260864,
1251
- 0.22378601133823395,
1252
- 0.23867575824260712,
1253
- 0.19821251928806305,
1254
- 0.2455597221851349,
1255
- 0.2343447506427765
1256
  ],
1257
  "base_loss": [
1258
- 11.928034782409668,
1259
- 11.653544425964355,
1260
- 11.307957649230957,
1261
- 10.721878051757812,
1262
- 10.181405067443848,
1263
- 9.667264938354492,
1264
- 9.305678367614746,
1265
- 9.201789855957031,
1266
- 8.780132293701172,
1267
- 8.337949752807617,
1268
- 8.104902267456055,
1269
- 7.672562599182129,
1270
- 7.409525394439697,
1271
- 6.95243501663208,
1272
- 6.568300724029541,
1273
- 6.088183879852295,
1274
- 5.540016174316406,
1275
- 5.10260534286499,
1276
- 4.484571933746338,
1277
- 4.0001749992370605,
1278
- 3.444486141204834,
1279
- 3.0357420444488525,
1280
- 2.7352375984191895,
1281
- 2.3701672554016113,
1282
- 2.2287418842315674,
1283
- 2.0617270469665527,
1284
- 2.0463647842407227,
1285
- 1.8895787000656128,
1286
- 1.8856637477874756,
1287
- 1.9227359294891357,
1288
- 1.8609199523925781,
1289
- 1.9265002012252808,
1290
- 1.9278079271316528,
1291
- 1.8539155721664429,
1292
- 1.8715951442718506,
1293
- 1.7844884395599365,
1294
- 1.8312513828277588,
1295
- 1.7931902408599854,
1296
- 1.810530424118042,
1297
- 1.8436228036880493,
1298
- 1.8527024984359741,
1299
- 1.8303594589233398,
1300
- 1.7831075191497803,
1301
- 1.7217122316360474,
1302
- 1.7545881271362305,
1303
- 1.7245683670043945,
1304
- 1.6838029623031616,
1305
- 1.7725518941879272,
1306
- 1.7447564601898193,
1307
- 1.6834079027175903,
1308
- 1.7119642496109009,
1309
- 1.7137107849121094,
1310
- 1.7549256086349487,
1311
- 1.7131061553955078,
1312
- 1.7283432483673096,
1313
- 1.7024879455566406,
1314
- 1.6835523843765259,
1315
- 1.724271535873413,
1316
- 1.6476503610610962,
1317
- 1.6461162567138672,
1318
- 1.6732215881347656,
1319
- 1.6631882190704346,
1320
- 1.6530768871307373,
1321
- 1.640184998512268,
1322
- 1.5616495609283447,
1323
- 1.5732653141021729,
1324
- 1.5028226375579834,
1325
- 1.469290852546692,
1326
- 1.4607034921646118,
1327
- 1.4396305084228516,
1328
- 1.33017897605896,
1329
- 1.173120141029358,
1330
- 1.0654069185256958,
1331
- 1.1612730026245117,
1332
- 1.1505136489868164,
1333
- 1.0570982694625854,
1334
- 0.9696107506752014,
1335
- 0.9057245254516602,
1336
- 0.8982640504837036,
1337
- 0.8462095856666565,
1338
- 0.7811938524246216,
1339
- 0.728919506072998,
1340
- 0.7578886151313782,
1341
- 0.7812106013298035,
1342
- 0.7021167874336243,
1343
- 0.7130362391471863,
1344
- 0.6820688247680664,
1345
- 0.6510214805603027,
1346
- 0.6453285217285156,
1347
- 0.6750946640968323,
1348
- 0.6662521362304688,
1349
- 0.589103102684021,
1350
- 0.635167121887207,
1351
- 0.6322607398033142,
1352
- 0.6354484558105469,
1353
- 0.5717305541038513,
1354
- 0.5770869851112366,
1355
- 0.5638580322265625,
1356
- 0.6003702282905579,
1357
- 0.606711208820343,
1358
- 0.5978370308876038,
1359
- 0.5625749826431274,
1360
- 0.5271865725517273,
1361
- 0.5920223593711853,
1362
- 0.6064496636390686,
1363
- 0.5727962255477905,
1364
- 0.5373821258544922,
1365
- 0.547116219997406,
1366
- 0.5527753233909607,
1367
- 0.5407260060310364,
1368
- 0.5330578088760376,
1369
- 0.5460835695266724,
1370
- 0.5407729744911194,
1371
- 0.5695903301239014,
1372
- 0.546602725982666,
1373
- 0.5379611253738403,
1374
- 0.49000412225723267,
1375
- 0.5377251505851746,
1376
- 0.5077806711196899,
1377
- 0.5233360528945923,
1378
- 0.5038724541664124,
1379
- 0.5082470774650574,
1380
- 0.5090785622596741,
1381
- 0.4877682328224182,
1382
- 0.46342095732688904,
1383
- 0.4934587776660919,
1384
- 0.47618719935417175,
1385
- 0.5213378071784973,
1386
- 0.48468828201293945,
1387
- 0.4764997959136963,
1388
- 0.490345299243927,
1389
- 0.48699042201042175,
1390
- 0.44104668498039246,
1391
- 0.474947452545166,
1392
- 0.46392837166786194,
1393
- 0.4587855041027069,
1394
- 0.4935186207294464,
1395
- 0.46145424246788025,
1396
- 0.4339943826198578,
1397
- 0.4798485338687897,
1398
- 0.4835148751735687,
1399
- 0.45113644003868103,
1400
- 0.4499821066856384,
1401
- 0.44402727484703064,
1402
- 0.4216662049293518,
1403
- 0.4317099452018738,
1404
- 0.4295271337032318,
1405
- 0.43690910935401917,
1406
- 0.4310178756713867,
1407
- 0.42988887429237366,
1408
- 0.39782506227493286,
1409
- 0.40841910243034363,
1410
- 0.4422263205051422,
1411
- 0.4496307075023651,
1412
- 0.4485880434513092,
1413
- 0.43126311898231506,
1414
- 0.41150495409965515,
1415
- 0.4591612219810486,
1416
- 0.45604208111763,
1417
- 0.46709761023521423,
1418
- 0.4348989427089691,
1419
- 0.3834414482116699,
1420
- 0.38175979256629944,
1421
- 0.3879620134830475,
1422
- 0.40807420015335083,
1423
- 0.4096015989780426,
1424
- 0.4135327637195587,
1425
- 0.3968922793865204,
1426
- 0.43453529477119446,
1427
- 0.36617913842201233,
1428
- 0.41760990023612976,
1429
- 0.3654819130897522,
1430
- 0.43267178535461426,
1431
- 0.3825819492340088,
1432
- 0.40239834785461426,
1433
- 0.4138084053993225,
1434
- 0.39564505219459534,
1435
- 0.36791282892227173,
1436
- 0.43040919303894043,
1437
- 0.4024931490421295,
1438
- 0.3815248906612396,
1439
- 0.40737414360046387,
1440
- 0.42400693893432617,
1441
- 0.38643333315849304,
1442
- 0.3655295670032501,
1443
- 0.3794443905353546,
1444
- 0.3761146664619446,
1445
- 0.402232825756073,
1446
- 0.39684897661209106,
1447
- 0.41486597061157227,
1448
- 0.3713434338569641,
1449
- 0.4122335612773895,
1450
- 0.3593781888484955,
1451
- 0.36669155955314636,
1452
- 0.38098907470703125,
1453
- 0.3585493564605713,
1454
- 0.3658697009086609,
1455
- 0.3952808976173401,
1456
- 0.3677598536014557,
1457
- 0.40938600897789,
1458
- 0.3865565359592438,
1459
- 0.3737422823905945,
1460
- 0.36394765973091125,
1461
- 0.33259841799736023,
1462
- 0.3456837832927704,
1463
- 0.33143261075019836,
1464
- 0.35574594140052795,
1465
- 0.3818463683128357,
1466
- 0.32026952505111694,
1467
- 0.35572943091392517,
1468
- 0.3279140889644623,
1469
- 0.38622188568115234,
1470
- 0.32534393668174744,
1471
- 0.35246992111206055,
1472
- 0.34478285908699036,
1473
- 0.3526977002620697,
1474
- 0.3618982434272766,
1475
- 0.3712795674800873,
1476
- 0.3642587661743164,
1477
- 0.3191487491130829,
1478
- 0.36171674728393555,
1479
- 0.34270623326301575,
1480
- 0.3760503828525543,
1481
- 0.34943634271621704,
1482
- 0.3346637785434723,
1483
- 0.32916831970214844,
1484
- 0.3213987350463867,
1485
- 0.34669747948646545,
1486
- 0.34980323910713196,
1487
- 0.3436082899570465,
1488
- 0.3271951675415039,
1489
- 0.31385931372642517,
1490
- 0.3484307825565338,
1491
- 0.3084122836589813,
1492
- 0.3468055725097656,
1493
- 0.3438337743282318,
1494
- 0.3528638482093811,
1495
- 0.31498146057128906,
1496
- 0.3160567879676819,
1497
- 0.34172186255455017,
1498
- 0.3417809009552002,
1499
- 0.3534678816795349,
1500
- 0.35489383339881897,
1501
- 0.34530016779899597,
1502
- 0.2865831255912781,
1503
- 0.3117128908634186,
1504
- 0.34365028142929077,
1505
- 0.33491435647010803,
1506
- 0.3689354360103607,
1507
- 0.3382095396518707,
1508
- 0.3239511549472809,
1509
- 0.3418653905391693,
1510
- 0.3021024167537689,
1511
- 0.2963162362575531,
1512
- 0.32997259497642517,
1513
- 0.33671286702156067,
1514
- 0.33611273765563965,
1515
- 0.3459137976169586,
1516
- 0.33286041021347046,
1517
- 0.2873530983924866,
1518
- 0.31002670526504517,
1519
- 0.30808526277542114,
1520
- 0.287386417388916,
1521
- 0.30139049887657166,
1522
- 0.29464051127433777,
1523
- 0.3082091510295868,
1524
- 0.3462357223033905,
1525
- 0.34610894322395325,
1526
- 0.3279619514942169,
1527
- 0.3180283010005951,
1528
- 0.2845425009727478,
1529
- 0.27627280354499817,
1530
- 0.29058805108070374,
1531
- 0.25387290120124817,
1532
- 0.29174312949180603,
1533
- 0.3285600244998932,
1534
- 0.31100231409072876,
1535
- 0.3193957507610321,
1536
- 0.3087748885154724,
1537
- 0.3061353862285614,
1538
- 0.32200178503990173,
1539
- 0.2894422709941864,
1540
- 0.2870963513851166,
1541
- 0.31858688592910767,
1542
- 0.2746082842350006,
1543
- 0.3215934932231903,
1544
- 0.3257999122142792,
1545
- 0.31921395659446716,
1546
- 0.2667975425720215,
1547
- 0.3186761736869812,
1548
- 0.2847942113876343,
1549
- 0.28523775935173035,
1550
- 0.32166609168052673,
1551
- 0.3076777458190918,
1552
- 0.3031381666660309,
1553
- 0.29915809631347656,
1554
- 0.32860395312309265,
1555
- 0.3052523136138916,
1556
- 0.31743431091308594,
1557
- 0.27312588691711426,
1558
- 0.26801934838294983,
1559
- 0.26507896184921265,
1560
- 0.294921338558197,
1561
- 0.3115091919898987,
1562
- 0.26985788345336914,
1563
- 0.293649286031723,
1564
- 0.2707940638065338,
1565
- 0.29856735467910767,
1566
- 0.2623736262321472,
1567
- 0.3006415367126465,
1568
- 0.31037160754203796,
1569
- 0.30440378189086914,
1570
- 0.3023694157600403,
1571
- 0.2869836390018463,
1572
- 0.24748651683330536,
1573
- 0.28105297684669495,
1574
- 0.2706967294216156,
1575
- 0.3016051948070526,
1576
- 0.2754139006137848,
1577
- 0.28681060671806335,
1578
- 0.28182414174079895,
1579
- 0.29593154788017273,
1580
- 0.2871556580066681,
1581
- 0.29613766074180603,
1582
- 0.30644792318344116,
1583
- 0.25587597489356995,
1584
- 0.28945502638816833,
1585
- 0.27285969257354736,
1586
- 0.28049322962760925,
1587
- 0.2884880602359772,
1588
- 0.2916378378868103,
1589
- 0.2754189372062683,
1590
- 0.3063291907310486,
1591
- 0.3094882071018219,
1592
- 0.29223495721817017,
1593
- 0.2699856460094452,
1594
- 0.265081524848938,
1595
- 0.3157065212726593,
1596
- 0.2825828790664673,
1597
- 0.2644478380680084,
1598
- 0.2736837565898895,
1599
- 0.2939130365848541,
1600
- 0.27335861325263977,
1601
- 0.27167394757270813,
1602
- 0.2819110155105591,
1603
- 0.289447158575058,
1604
- 0.2571549415588379,
1605
- 0.2593221962451935,
1606
- 0.28376683592796326,
1607
- 0.29852765798568726,
1608
- 0.27658429741859436,
1609
- 0.23924340307712555,
1610
- 0.28641754388809204,
1611
- 0.3209593594074249,
1612
- 0.2669186592102051,
1613
- 0.2522584795951843,
1614
- 0.2631411552429199,
1615
- 0.30004528164863586,
1616
- 0.25664424896240234,
1617
- 0.26312512159347534,
1618
- 0.2851202189922333,
1619
- 0.2921363413333893,
1620
- 0.24688827991485596,
1621
- 0.2517416179180145,
1622
- 0.30236250162124634,
1623
- 0.24213668704032898,
1624
- 0.2524048686027527,
1625
- 0.27556005120277405,
1626
- 0.27097469568252563,
1627
- 0.29630640149116516,
1628
- 0.24852998554706573,
1629
- 0.29177477955818176,
1630
- 0.2854974567890167,
1631
- 0.2374695986509323,
1632
- 0.23592327535152435,
1633
- 0.3132108151912689,
1634
- 0.2631916403770447,
1635
- 0.2837892472743988,
1636
- 0.25606682896614075,
1637
- 0.27506956458091736,
1638
- 0.24950344860553741,
1639
- 0.2695748507976532,
1640
- 0.2633344829082489,
1641
- 0.2621081471443176,
1642
- 0.2588253915309906,
1643
- 0.26628708839416504,
1644
- 0.2546164393424988,
1645
- 0.2710006535053253,
1646
- 0.2939824163913727,
1647
- 0.23946473002433777,
1648
- 0.22712087631225586,
1649
- 0.2623310983181,
1650
- 0.2482786327600479,
1651
- 0.27607423067092896,
1652
- 0.25139644742012024,
1653
- 0.28528496623039246,
1654
- 0.24713945388793945,
1655
- 0.24412168562412262,
1656
- 0.25628599524497986,
1657
- 0.28033557534217834,
1658
- 0.26747894287109375,
1659
- 0.29976686835289,
1660
- 0.29141291975975037,
1661
- 0.2702557146549225,
1662
- 0.24962809681892395,
1663
- 0.2612250745296478,
1664
- 0.2625961899757385,
1665
- 0.2644568383693695,
1666
- 0.2803192436695099,
1667
- 0.23608256876468658,
1668
- 0.23363173007965088,
1669
- 0.2611221969127655,
1670
- 0.2525368332862854,
1671
- 0.2524564564228058,
1672
- 0.24044832587242126,
1673
- 0.2552793622016907,
1674
- 0.28079238533973694,
1675
- 0.2619774639606476,
1676
- 0.23810975253582,
1677
- 0.24849574267864227,
1678
- 0.25039538741111755,
1679
- 0.2613367438316345,
1680
- 0.2615041732788086,
1681
- 0.20929165184497833,
1682
- 0.23789460957050323,
1683
- 0.2508893311023712,
1684
- 0.23562470078468323,
1685
- 0.27942875027656555,
1686
- 0.282475084066391,
1687
- 0.26715087890625,
1688
- 0.25968289375305176,
1689
- 0.26287397742271423,
1690
- 0.25527840852737427,
1691
- 0.24361786246299744,
1692
- 0.24610944092273712,
1693
- 0.2573030889034271,
1694
- 0.28686219453811646,
1695
- 0.25092869997024536,
1696
- 0.26785707473754883,
1697
- 0.2560536861419678,
1698
- 0.2647089958190918,
1699
- 0.2567795217037201,
1700
- 0.2601569592952728,
1701
- 0.24640819430351257,
1702
- 0.2674255669116974,
1703
- 0.2340751439332962,
1704
- 0.22037099301815033,
1705
- 0.23103085160255432,
1706
- 0.22777055203914642,
1707
- 0.24295179545879364,
1708
- 0.23772236704826355,
1709
- 0.23818251490592957,
1710
- 0.24417351186275482,
1711
- 0.23382483422756195,
1712
- 0.2717232406139374,
1713
- 0.24655917286872864,
1714
- 0.24539153277873993,
1715
- 0.27003398537635803,
1716
- 0.24944210052490234,
1717
- 0.21866385638713837,
1718
- 0.24142920970916748,
1719
- 0.2841137945652008,
1720
- 0.27210989594459534,
1721
- 0.24623851478099823,
1722
- 0.2414173036813736,
1723
- 0.24330724775791168,
1724
- 0.2612263560295105,
1725
- 0.24263326823711395,
1726
- 0.23556897044181824,
1727
- 0.25661733746528625,
1728
- 0.23128819465637207,
1729
- 0.2515091300010681,
1730
- 0.23440249264240265,
1731
- 0.26295438408851624,
1732
- 0.24453571438789368,
1733
- 0.25462159514427185,
1734
- 0.23500898480415344,
1735
- 0.24620413780212402,
1736
- 0.23643255233764648,
1737
- 0.21674731373786926,
1738
- 0.23458817601203918,
1739
- 0.22282226383686066,
1740
- 0.2536497414112091,
1741
- 0.23018422722816467,
1742
- 0.21616289019584656,
1743
- 0.24523791670799255,
1744
- 0.2539035975933075,
1745
- 0.2417120784521103,
1746
- 0.2687821388244629,
1747
- 0.2591947615146637,
1748
- 0.23294435441493988,
1749
- 0.24325568974018097,
1750
- 0.2612447142601013,
1751
- 0.23927579820156097,
1752
- 0.23915018141269684,
1753
- 0.26768800616264343,
1754
- 0.24271290004253387,
1755
- 0.25881507992744446,
1756
- 0.23990657925605774,
1757
- 0.22135673463344574,
1758
- 0.20841874182224274,
1759
- 0.24352209270000458,
1760
- 0.2426811009645462,
1761
- 0.22436712682247162,
1762
- 0.24585120379924774,
1763
- 0.23543156683444977,
1764
- 0.24074861407279968,
1765
- 0.2525044083595276,
1766
- 0.24585461616516113,
1767
- 0.22054941952228546,
1768
- 0.23277811706066132,
1769
- 0.2542433738708496,
1770
- 0.21179088950157166,
1771
- 0.2887387275695801,
1772
- 0.24148423969745636,
1773
- 0.25414398312568665,
1774
- 0.25563937425613403,
1775
- 0.23944461345672607,
1776
- 0.2533447742462158,
1777
- 0.23563027381896973,
1778
- 0.26421570777893066,
1779
- 0.2254643440246582,
1780
- 0.24568693339824677,
1781
- 0.2360592931509018,
1782
- 0.2221137136220932,
1783
- 0.2282101809978485,
1784
- 0.24913156032562256,
1785
- 0.22136715054512024,
1786
- 0.23910680413246155,
1787
- 0.23922684788703918,
1788
- 0.22790290415287018,
1789
- 0.23584823310375214,
1790
- 0.2730078399181366,
1791
- 0.2663761079311371,
1792
- 0.248269185423851,
1793
- 0.25979501008987427,
1794
- 0.2336592972278595,
1795
- 0.2624785006046295,
1796
- 0.22696606814861298,
1797
- 0.21379384398460388,
1798
- 0.21497604250907898,
1799
- 0.2592029571533203,
1800
- 0.22469215095043182,
1801
- 0.25158485770225525,
1802
- 0.24422571063041687,
1803
- 0.25753626227378845,
1804
- 0.25007328391075134,
1805
- 0.2602040469646454,
1806
- 0.23332370817661285,
1807
- 0.23722335696220398,
1808
- 0.24797333776950836,
1809
- 0.22556526958942413,
1810
- 0.23421958088874817,
1811
- 0.24832820892333984,
1812
- 0.2311556339263916,
1813
- 0.23707018792629242,
1814
- 0.23783321678638458,
1815
- 0.20987555384635925,
1816
- 0.267695814371109,
1817
- 0.2629874646663666,
1818
- 0.21873648464679718,
1819
- 0.24494552612304688,
1820
- 0.26303473114967346,
1821
- 0.24542763829231262,
1822
- 0.23399154841899872,
1823
- 0.23643358051776886,
1824
- 0.21889302134513855,
1825
- 0.22333727777004242,
1826
- 0.20176778733730316,
1827
- 0.2578296363353729,
1828
- 0.26290076971054077,
1829
- 0.2378210574388504,
1830
- 0.24513255059719086,
1831
- 0.25607389211654663,
1832
- 0.24997523427009583,
1833
- 0.2558247745037079,
1834
- 0.25033673644065857,
1835
- 0.2546168267726898,
1836
- 0.23066361248493195,
1837
- 0.22089658677577972,
1838
- 0.2460862696170807,
1839
- 0.21666987240314484,
1840
- 0.22690625488758087,
1841
- 0.24581076204776764,
1842
- 0.2427884191274643,
1843
- 0.22972403466701508,
1844
- 0.25537118315696716,
1845
- 0.24844148755073547,
1846
- 0.2209678441286087,
1847
- 0.22279927134513855,
1848
- 0.23022934794425964,
1849
- 0.2374199479818344,
1850
- 0.2532387375831604,
1851
- 0.24079224467277527,
1852
- 0.2286282330751419,
1853
- 0.22575612366199493,
1854
- 0.22752560675144196,
1855
- 0.26504701375961304,
1856
- 0.23773540556430817,
1857
- 0.20410892367362976,
1858
- 0.25362756848335266,
1859
- 0.24984030425548553,
1860
- 0.2467348873615265,
1861
- 0.2259712964296341,
1862
- 0.23826143145561218,
1863
- 0.23323170840740204,
1864
- 0.23948879539966583,
1865
- 0.24911201000213623,
1866
- 0.23734129965305328,
1867
- 0.24040012061595917,
1868
- 0.26085561513900757,
1869
- 0.25282594561576843,
1870
- 0.24620071053504944,
1871
- 0.2408737689256668,
1872
- 0.22477446496486664,
1873
- 0.232040137052536,
1874
- 0.22941991686820984,
1875
- 0.2574242055416107,
1876
- 0.22758789360523224,
1877
- 0.2541619539260864,
1878
- 0.22378601133823395,
1879
- 0.23867575824260712,
1880
- 0.19821251928806305,
1881
- 0.2455597221851349,
1882
- 0.2343447506427765
1883
  ],
1884
  "lr": [
1885
  1.0458911419423694e-06,
@@ -2553,595 +2553,595 @@
2553
  20
2554
  ],
2555
  "eval_accuracy": [
2556
- 0.014444444444444444,
2557
- 0.0044444444444444444,
2558
- 0.08666666666666667,
2559
- 0.08666666666666667,
2560
- 0.2011111111111111,
2561
- 0.24555555555555555,
2562
- 0.27555555555555555,
2563
- 0.2577777777777778,
2564
- 0.33,
2565
- 0.3088888888888889,
2566
- 0.35,
2567
- 0.34,
2568
- 0.3611111111111111,
2569
- 0.3988888888888889,
2570
- 0.39222222222222225,
2571
- 0.39666666666666667,
2572
- 0.41,
2573
- 0.4122222222222222,
2574
- 0.4022222222222222,
2575
- 0.4022222222222222
2576
  ]
2577
  },
2578
- "final_accuracy": 0.33375,
2579
  "sft_eval": {
2580
  "config": {
2581
  "ops": "add_sub",
2582
  "K": null,
2583
  "mode": "sft",
2584
  "n_digits": 6,
2585
- "n_per_split": 50
2586
  },
2587
  "splits": {
2588
  "add_S0": {
2589
- "full_accuracy": 0.68,
2590
- "digit_accuracy": 0.9485714285714286,
2591
- "n_examples": 50,
2592
  "per_subtask": {
2593
  "SA": {
2594
- "accuracy": 0.9457627118644067,
2595
- "count": 295
2596
  },
2597
  "SS": {
2598
- "accuracy": 0.9636363636363636,
2599
- "count": 55
2600
  }
2601
  }
2602
  },
2603
  "add_S1": {
2604
- "full_accuracy": 0.76,
2605
- "digit_accuracy": 0.9542857142857143,
2606
- "n_examples": 50,
2607
  "per_subtask": {
2608
  "SA": {
2609
- "accuracy": 0.9761904761904762,
2610
- "count": 126
2611
  },
2612
  "SC": {
2613
- "accuracy": 0.9240506329113924,
2614
- "count": 79
2615
  },
2616
  "SS": {
2617
- "accuracy": 1.0,
2618
- "count": 21
2619
  },
2620
  "UC": {
2621
- "accuracy": 0.9435483870967742,
2622
- "count": 124
2623
  }
2624
  }
2625
  },
2626
  "add_S2": {
2627
- "full_accuracy": 0.32,
2628
- "digit_accuracy": 0.8542857142857143,
2629
- "n_examples": 50,
2630
  "per_subtask": {
2631
  "SA": {
2632
- "accuracy": 0.9466666666666667,
2633
- "count": 75
2634
  },
2635
  "SC": {
2636
- "accuracy": 0.8548387096774194,
2637
- "count": 62
2638
  },
2639
  "SS": {
2640
- "accuracy": 0.8205128205128205,
2641
- "count": 39
2642
  },
2643
  "UC": {
2644
- "accuracy": 0.7747747747747747,
2645
- "count": 111
2646
  },
2647
  "US": {
2648
- "accuracy": 0.9047619047619048,
2649
- "count": 63
2650
  }
2651
  }
2652
  },
2653
  "add_S3": {
2654
  "full_accuracy": 0.18,
2655
- "digit_accuracy": 0.7857142857142857,
2656
- "n_examples": 50,
2657
  "per_subtask": {
2658
  "SA": {
2659
- "accuracy": 0.9333333333333333,
2660
- "count": 60
2661
  },
2662
  "SC": {
2663
- "accuracy": 0.9298245614035088,
2664
- "count": 57
2665
  },
2666
  "SS": {
2667
- "accuracy": 0.8947368421052632,
2668
- "count": 19
2669
  },
2670
  "UC": {
2671
- "accuracy": 0.6923076923076923,
2672
- "count": 104
2673
  },
2674
  "US": {
2675
- "accuracy": 0.7,
2676
- "count": 110
2677
  }
2678
  }
2679
  },
2680
  "add_S4": {
2681
- "full_accuracy": 0.14,
2682
- "digit_accuracy": 0.6371428571428571,
2683
- "n_examples": 50,
2684
  "per_subtask": {
2685
  "SA": {
2686
- "accuracy": 1.0,
2687
- "count": 48
2688
  },
2689
  "SC": {
2690
- "accuracy": 0.9423076923076923,
2691
- "count": 52
2692
  },
2693
  "SS": {
2694
- "accuracy": 0.8571428571428571,
2695
- "count": 7
2696
  },
2697
  "UC": {
2698
- "accuracy": 0.5168539325842697,
2699
- "count": 89
2700
  },
2701
  "US": {
2702
- "accuracy": 0.4805194805194805,
2703
- "count": 154
2704
  }
2705
  }
2706
  },
2707
  "add_S5": {
2708
- "full_accuracy": 0.08,
2709
- "digit_accuracy": 0.4257142857142857,
2710
- "n_examples": 50,
2711
  "per_subtask": {
2712
  "SA": {
2713
  "accuracy": 1.0,
2714
- "count": 50
2715
  },
2716
  "SC": {
2717
- "accuracy": 1.0,
2718
- "count": 50
2719
  },
2720
  "UC": {
2721
- "accuracy": 0.24,
2722
- "count": 50
2723
  },
2724
  "US": {
2725
- "accuracy": 0.185,
2726
- "count": 200
2727
  }
2728
  }
2729
  },
2730
  "add_S6": {
2731
- "full_accuracy": 0.4,
2732
- "digit_accuracy": 0.52,
2733
- "n_examples": 50,
2734
  "per_subtask": {
2735
  "SC": {
2736
  "accuracy": 1.0,
2737
- "count": 50
2738
  },
2739
  "UC": {
2740
- "accuracy": 0.4,
2741
- "count": 50
2742
  },
2743
  "US": {
2744
- "accuracy": 0.448,
2745
- "count": 250
2746
  }
2747
  }
2748
  },
2749
  "add_random": {
2750
- "full_accuracy": 0.62,
2751
- "digit_accuracy": 0.9321428571428572,
2752
  "n_examples": 200,
2753
  "per_subtask": {
2754
  "SA": {
2755
- "accuracy": 0.9767981438515081,
2756
- "count": 431
2757
  },
2758
  "SC": {
2759
- "accuracy": 0.9588607594936709,
2760
- "count": 316
2761
  },
2762
  "SS": {
2763
- "accuracy": 1.0,
2764
- "count": 39
2765
  },
2766
  "UC": {
2767
- "accuracy": 0.8892857142857142,
2768
- "count": 560
2769
  },
2770
  "US": {
2771
- "accuracy": 0.8148148148148148,
2772
- "count": 54
2773
  }
2774
  }
2775
  },
2776
  "add_C1": {
2777
- "full_accuracy": 0.86,
2778
- "digit_accuracy": 0.98,
2779
- "n_examples": 50,
2780
  "per_subtask": {
2781
  "SA": {
2782
- "accuracy": 0.992,
2783
- "count": 250
2784
  },
2785
  "SC": {
2786
- "accuracy": 0.98,
2787
- "count": 50
2788
  },
2789
  "UC": {
2790
- "accuracy": 0.92,
2791
- "count": 50
2792
  }
2793
  }
2794
  },
2795
  "add_C2": {
2796
- "full_accuracy": 0.66,
2797
- "digit_accuracy": 0.9485714285714286,
2798
- "n_examples": 50,
2799
  "per_subtask": {
2800
  "SA": {
2801
- "accuracy": 0.985,
2802
- "count": 200
2803
  },
2804
  "SC": {
2805
- "accuracy": 1.0,
2806
- "count": 50
2807
  },
2808
  "UC": {
2809
- "accuracy": 0.8433734939759037,
2810
- "count": 83
2811
  },
2812
  "US": {
2813
- "accuracy": 0.8823529411764706,
2814
- "count": 17
2815
  }
2816
  }
2817
  },
2818
  "add_C3": {
2819
- "full_accuracy": 0.36,
2820
- "digit_accuracy": 0.8628571428571429,
2821
- "n_examples": 50,
2822
  "per_subtask": {
2823
  "SA": {
2824
- "accuracy": 0.9933333333333333,
2825
- "count": 150
2826
  },
2827
  "SC": {
2828
- "accuracy": 1.0,
2829
- "count": 50
2830
  },
2831
  "UC": {
2832
- "accuracy": 0.68,
2833
- "count": 100
2834
  },
2835
  "US": {
2836
- "accuracy": 0.7,
2837
- "count": 50
2838
  }
2839
  }
2840
  },
2841
  "add_C4": {
2842
- "full_accuracy": 0.3,
2843
- "digit_accuracy": 0.8542857142857143,
2844
- "n_examples": 50,
2845
  "per_subtask": {
2846
  "SA": {
2847
- "accuracy": 0.99,
2848
- "count": 100
2849
  },
2850
  "SC": {
2851
  "accuracy": 1.0,
2852
- "count": 50
2853
  },
2854
  "UC": {
2855
- "accuracy": 0.7121212121212122,
2856
- "count": 132
2857
  },
2858
  "US": {
2859
- "accuracy": 0.8235294117647058,
2860
- "count": 68
2861
  }
2862
  }
2863
  },
2864
  "add_C5": {
2865
- "full_accuracy": 0.26,
2866
- "digit_accuracy": 0.7828571428571428,
2867
- "n_examples": 50,
2868
  "per_subtask": {
2869
  "SA": {
2870
  "accuracy": 1.0,
2871
- "count": 50
2872
  },
2873
  "SC": {
2874
- "accuracy": 1.0,
2875
- "count": 50
2876
  },
2877
  "UC": {
2878
- "accuracy": 0.7465753424657534,
2879
- "count": 146
2880
  },
2881
  "US": {
2882
- "accuracy": 0.625,
2883
- "count": 104
2884
  }
2885
  }
2886
  },
2887
  "add_C6": {
2888
- "full_accuracy": 0.3,
2889
- "digit_accuracy": 0.7971428571428572,
2890
- "n_examples": 50,
2891
  "per_subtask": {
2892
  "SC": {
2893
  "accuracy": 1.0,
2894
- "count": 50
2895
  },
2896
  "UC": {
2897
- "accuracy": 0.7724867724867724,
2898
- "count": 189
2899
  },
2900
  "US": {
2901
- "accuracy": 0.7477477477477478,
2902
- "count": 111
2903
  }
2904
  }
2905
  },
2906
  "sub_M0": {
2907
- "full_accuracy": 0.8,
2908
- "digit_accuracy": 0.9714285714285714,
2909
- "n_examples": 50,
2910
  "per_subtask": {
2911
  "MD": {
2912
- "accuracy": 0.966996699669967,
2913
- "count": 303
2914
  },
2915
  "ME": {
2916
- "accuracy": 1.0,
2917
- "count": 47
2918
  }
2919
  }
2920
  },
2921
  "sub_M1": {
2922
  "full_accuracy": 0.58,
2923
- "digit_accuracy": 0.9314285714285714,
2924
- "n_examples": 50,
2925
  "per_subtask": {
2926
  "MD": {
2927
- "accuracy": 0.9858156028368794,
2928
- "count": 141
2929
  },
2930
  "MB": {
2931
- "accuracy": 0.9305555555555556,
2932
- "count": 72
2933
  },
2934
  "ME": {
2935
  "accuracy": 1.0,
2936
- "count": 18
2937
  },
2938
  "UB": {
2939
- "accuracy": 0.8571428571428571,
2940
- "count": 119
2941
  }
2942
  }
2943
  },
2944
  "sub_M2": {
2945
- "full_accuracy": 0.14,
2946
- "digit_accuracy": 0.8228571428571428,
2947
- "n_examples": 50,
2948
  "per_subtask": {
2949
  "MD": {
2950
- "accuracy": 0.9553571428571429,
2951
- "count": 112
2952
  },
2953
  "MB": {
2954
- "accuracy": 0.8679245283018868,
2955
- "count": 53
2956
  },
2957
  "ME": {
2958
- "accuracy": 0.9787234042553191,
2959
- "count": 47
2960
  },
2961
  "UB": {
2962
- "accuracy": 0.5529411764705883,
2963
- "count": 85
2964
  },
2965
  "UD": {
2966
- "accuracy": 0.7924528301886793,
2967
- "count": 53
2968
  }
2969
  }
2970
  },
2971
  "sub_M3": {
2972
- "full_accuracy": 0.06,
2973
- "digit_accuracy": 0.7228571428571429,
2974
- "n_examples": 50,
2975
  "per_subtask": {
2976
  "MD": {
2977
- "accuracy": 0.9690721649484536,
2978
- "count": 97
2979
  },
2980
  "MB": {
2981
- "accuracy": 0.9803921568627451,
2982
- "count": 51
2983
  },
2984
  "ME": {
2985
- "accuracy": 0.9629629629629629,
2986
- "count": 27
2987
  },
2988
  "UB": {
2989
- "accuracy": 0.5405405405405406,
2990
- "count": 74
2991
  },
2992
  "UD": {
2993
- "accuracy": 0.42574257425742573,
2994
- "count": 101
2995
  }
2996
  }
2997
  },
2998
  "sub_M4": {
2999
- "full_accuracy": 0.06,
3000
- "digit_accuracy": 0.6228571428571429,
3001
- "n_examples": 50,
3002
  "per_subtask": {
3003
  "MD": {
3004
- "accuracy": 0.98,
3005
- "count": 100
3006
  },
3007
  "MB": {
3008
- "accuracy": 0.98,
3009
- "count": 50
3010
  },
3011
  "UB": {
3012
- "accuracy": 0.54,
3013
- "count": 50
3014
  },
3015
  "UD": {
3016
- "accuracy": 0.29333333333333333,
3017
- "count": 150
3018
  }
3019
  }
3020
  },
3021
  "sub_M5": {
3022
- "full_accuracy": 0.06,
3023
- "digit_accuracy": 0.5114285714285715,
3024
- "n_examples": 50,
3025
  "per_subtask": {
3026
  "MD": {
3027
  "accuracy": 1.0,
3028
- "count": 50
3029
  },
3030
  "MB": {
3031
  "accuracy": 1.0,
3032
- "count": 50
3033
  },
3034
  "UB": {
3035
- "accuracy": 0.38,
3036
- "count": 50
3037
  },
3038
  "UD": {
3039
- "accuracy": 0.3,
3040
- "count": 200
3041
  }
3042
  }
3043
  },
3044
  "sub_random": {
3045
- "full_accuracy": 0.525,
3046
- "digit_accuracy": 0.9178571428571428,
3047
  "n_examples": 200,
3048
  "per_subtask": {
3049
  "MD": {
3050
- "accuracy": 0.9736842105263158,
3051
- "count": 570
3052
  },
3053
  "MB": {
3054
- "accuracy": 0.9675090252707581,
3055
- "count": 277
3056
  },
3057
  "ME": {
3058
- "accuracy": 1.0,
3059
  "count": 53
3060
  },
3061
  "UB": {
3062
- "accuracy": 0.8131634819532909,
3063
- "count": 471
3064
  },
3065
  "UD": {
3066
- "accuracy": 0.896551724137931,
3067
- "count": 29
3068
  }
3069
  }
3070
  },
3071
  "sub_B3": {
3072
- "full_accuracy": 0.26,
3073
- "digit_accuracy": 0.8285714285714286,
3074
- "n_examples": 50,
3075
  "per_subtask": {
3076
  "MD": {
3077
- "accuracy": 0.9866666666666667,
3078
- "count": 150
3079
  },
3080
  "MB": {
3081
- "accuracy": 0.96,
3082
- "count": 50
3083
  },
3084
  "UB": {
3085
- "accuracy": 0.6237623762376238,
3086
- "count": 101
3087
  },
3088
  "UD": {
3089
- "accuracy": 0.6326530612244898,
3090
- "count": 49
3091
  }
3092
  }
3093
  },
3094
  "sub_B4": {
3095
- "full_accuracy": 0.1,
3096
- "digit_accuracy": 0.7371428571428571,
3097
- "n_examples": 50,
3098
  "per_subtask": {
3099
  "MD": {
3100
- "accuracy": 0.98,
3101
- "count": 100
3102
  },
3103
  "MB": {
3104
- "accuracy": 0.94,
3105
- "count": 50
3106
  },
3107
  "UB": {
3108
- "accuracy": 0.5867768595041323,
3109
- "count": 121
3110
  },
3111
  "UD": {
3112
- "accuracy": 0.5316455696202531,
3113
- "count": 79
3114
  }
3115
  }
3116
  },
3117
  "sub_B5": {
3118
- "full_accuracy": 0.16,
3119
- "digit_accuracy": 0.7028571428571428,
3120
- "n_examples": 50,
3121
  "per_subtask": {
3122
  "MD": {
3123
  "accuracy": 1.0,
3124
- "count": 50
3125
  },
3126
  "MB": {
3127
  "accuracy": 1.0,
3128
- "count": 50
3129
  },
3130
  "UB": {
3131
- "accuracy": 0.5986842105263158,
3132
- "count": 152
3133
  },
3134
  "UD": {
3135
- "accuracy": 0.5612244897959183,
3136
- "count": 98
3137
  }
3138
  }
3139
  }
3140
  },
3141
  "summary": {
3142
- "overall_accuracy": 0.4026666666666667,
3143
- "digit_accuracy": 0.82,
3144
- "total_examples": 1500,
3145
  "n_splits": 24
3146
  }
3147
  }
 
628
  31250
629
  ],
630
  "loss": [
631
+ 11.963690757751465,
632
+ 11.661938667297363,
633
+ 11.223055839538574,
634
+ 10.688573837280273,
635
+ 9.94072437286377,
636
+ 9.775545120239258,
637
+ 9.39625072479248,
638
+ 9.00570011138916,
639
+ 8.761700630187988,
640
+ 8.4192533493042,
641
+ 8.12822437286377,
642
+ 7.931671142578125,
643
+ 7.313607215881348,
644
+ 6.978346347808838,
645
+ 6.648111820220947,
646
+ 5.993403911590576,
647
+ 5.525230407714844,
648
+ 5.094429969787598,
649
+ 4.402842998504639,
650
+ 3.949738025665283,
651
+ 3.4472267627716064,
652
+ 2.9571051597595215,
653
+ 2.724224328994751,
654
+ 2.427931070327759,
655
+ 2.2689688205718994,
656
+ 2.1323626041412354,
657
+ 2.0063560009002686,
658
+ 2.0175435543060303,
659
+ 1.9365376234054565,
660
+ 1.9525864124298096,
661
+ 1.9771562814712524,
662
+ 1.9292795658111572,
663
+ 1.9087927341461182,
664
+ 1.8182607889175415,
665
+ 1.8655515909194946,
666
+ 1.7887706756591797,
667
+ 1.8116521835327148,
668
+ 1.8060662746429443,
669
+ 1.8420522212982178,
670
+ 1.8362396955490112,
671
+ 1.7543896436691284,
672
+ 1.7854712009429932,
673
+ 1.838714599609375,
674
+ 1.8220285177230835,
675
+ 1.824865460395813,
676
+ 1.744118332862854,
677
+ 1.7437489032745361,
678
+ 1.8406013250350952,
679
+ 1.769796371459961,
680
+ 1.7702674865722656,
681
+ 1.6466058492660522,
682
+ 1.8461577892303467,
683
+ 1.7179840803146362,
684
+ 1.6748616695404053,
685
+ 1.6761716604232788,
686
+ 1.7217031717300415,
687
+ 1.7508471012115479,
688
+ 1.7157037258148193,
689
+ 1.6086606979370117,
690
+ 1.6273046731948853,
691
+ 1.6577644348144531,
692
+ 1.5764987468719482,
693
+ 1.6589974164962769,
694
+ 1.616567850112915,
695
+ 1.6576578617095947,
696
+ 1.4999885559082031,
697
+ 1.5213489532470703,
698
+ 1.5925500392913818,
699
+ 1.4423539638519287,
700
+ 1.528339147567749,
701
+ 1.402854323387146,
702
+ 1.3988912105560303,
703
+ 1.367125153541565,
704
+ 1.2756946086883545,
705
+ 1.2506844997406006,
706
+ 1.173041582107544,
707
+ 1.193954586982727,
708
+ 0.990950882434845,
709
+ 0.9718019366264343,
710
+ 0.943374752998352,
711
+ 0.943237841129303,
712
+ 0.8248871564865112,
713
+ 0.9327092170715332,
714
+ 0.7993831634521484,
715
+ 0.7706826329231262,
716
+ 0.7637476325035095,
717
+ 0.7574962973594666,
718
+ 0.7494969367980957,
719
+ 0.7162280678749084,
720
+ 0.7451773881912231,
721
+ 0.670974612236023,
722
+ 0.663203239440918,
723
+ 0.6701799035072327,
724
+ 0.6127704381942749,
725
+ 0.6420816779136658,
726
+ 0.6314144134521484,
727
+ 0.6287131905555725,
728
+ 0.5879684090614319,
729
+ 0.6163527369499207,
730
+ 0.598551869392395,
731
+ 0.6191786527633667,
732
+ 0.5709514617919922,
733
+ 0.5744446516036987,
734
+ 0.5953712463378906,
735
+ 0.5669698119163513,
736
+ 0.569784939289093,
737
+ 0.590087354183197,
738
+ 0.6043781042098999,
739
+ 0.5754431486129761,
740
+ 0.5330023169517517,
741
+ 0.5477746725082397,
742
+ 0.5358116626739502,
743
+ 0.5499526262283325,
744
+ 0.5191396474838257,
745
+ 0.5597048997879028,
746
+ 0.5326811671257019,
747
+ 0.5370150804519653,
748
+ 0.547336995601654,
749
+ 0.5351187586784363,
750
+ 0.5349653959274292,
751
+ 0.5326358079910278,
752
+ 0.5140055418014526,
753
+ 0.5046959519386292,
754
+ 0.49848243594169617,
755
+ 0.5145978927612305,
756
+ 0.489215224981308,
757
+ 0.5175383687019348,
758
+ 0.48192352056503296,
759
+ 0.5176753401756287,
760
+ 0.4990863502025604,
761
+ 0.49212411046028137,
762
+ 0.5086663365364075,
763
+ 0.4860772490501404,
764
+ 0.5174476504325867,
765
+ 0.5017643570899963,
766
+ 0.4740331470966339,
767
+ 0.46239525079727173,
768
+ 0.4612799286842346,
769
+ 0.4830116629600525,
770
+ 0.464834064245224,
771
+ 0.4621468186378479,
772
+ 0.46891137957572937,
773
+ 0.5075568556785583,
774
+ 0.43174198269844055,
775
+ 0.46206027269363403,
776
+ 0.4287765920162201,
777
+ 0.4551263451576233,
778
+ 0.44393548369407654,
779
+ 0.43048691749572754,
780
+ 0.4558803141117096,
781
+ 0.4532429873943329,
782
+ 0.4646574556827545,
783
+ 0.431844025850296,
784
+ 0.4672628939151764,
785
+ 0.44547510147094727,
786
+ 0.4079582691192627,
787
+ 0.4426649212837219,
788
+ 0.39009925723075867,
789
+ 0.43490490317344666,
790
+ 0.44311219453811646,
791
+ 0.39784926176071167,
792
+ 0.41195815801620483,
793
+ 0.42954376339912415,
794
+ 0.41115802526474,
795
+ 0.4012797176837921,
796
+ 0.38349026441574097,
797
+ 0.43120235204696655,
798
+ 0.39938807487487793,
799
+ 0.4433351457118988,
800
+ 0.4032042324542999,
801
+ 0.4240035116672516,
802
+ 0.4189927577972412,
803
+ 0.3758549094200134,
804
+ 0.4081072211265564,
805
+ 0.41931289434432983,
806
+ 0.396896094083786,
807
+ 0.37466028332710266,
808
+ 0.41262954473495483,
809
+ 0.3888600766658783,
810
+ 0.38536521792411804,
811
+ 0.4092770516872406,
812
+ 0.3643164336681366,
813
+ 0.3778727054595947,
814
+ 0.3487100303173065,
815
+ 0.38028067350387573,
816
+ 0.36624351143836975,
817
+ 0.3879108428955078,
818
+ 0.3812934458255768,
819
+ 0.4147397577762604,
820
+ 0.32763606309890747,
821
+ 0.362922728061676,
822
+ 0.37037578225135803,
823
+ 0.3684242069721222,
824
+ 0.3541705906391144,
825
+ 0.3531418442726135,
826
+ 0.38803714513778687,
827
+ 0.3663487434387207,
828
+ 0.34989604353904724,
829
+ 0.3375271260738373,
830
+ 0.3937970995903015,
831
+ 0.36703363060951233,
832
+ 0.36325493454933167,
833
+ 0.35152706503868103,
834
+ 0.36929401755332947,
835
+ 0.3409608006477356,
836
+ 0.3745754659175873,
837
+ 0.36702921986579895,
838
+ 0.36814290285110474,
839
+ 0.34607431292533875,
840
+ 0.3489746153354645,
841
+ 0.3475476801395416,
842
+ 0.3784462809562683,
843
+ 0.37143474817276,
844
+ 0.3513146936893463,
845
+ 0.3392271101474762,
846
+ 0.37690967321395874,
847
+ 0.35800108313560486,
848
+ 0.35400205850601196,
849
+ 0.3127591609954834,
850
+ 0.3481731116771698,
851
+ 0.3677438199520111,
852
+ 0.37050268054008484,
853
+ 0.3325396180152893,
854
+ 0.33604928851127625,
855
+ 0.34623295068740845,
856
+ 0.32801350951194763,
857
+ 0.3004973828792572,
858
+ 0.33102330565452576,
859
+ 0.35216864943504333,
860
+ 0.35869526863098145,
861
+ 0.3230397701263428,
862
+ 0.34125009179115295,
863
+ 0.3288026750087738,
864
+ 0.36230114102363586,
865
+ 0.32093361020088196,
866
+ 0.33653420209884644,
867
+ 0.31276172399520874,
868
+ 0.32386818528175354,
869
+ 0.30667203664779663,
870
+ 0.3109426200389862,
871
+ 0.3534684181213379,
872
+ 0.3197288513183594,
873
+ 0.30439478158950806,
874
+ 0.292362242937088,
875
+ 0.3334762752056122,
876
+ 0.35194072127342224,
877
+ 0.29283004999160767,
878
+ 0.32437053322792053,
879
+ 0.293384850025177,
880
+ 0.31290170550346375,
881
+ 0.32321569323539734,
882
+ 0.32250460982322693,
883
+ 0.3221062123775482,
884
+ 0.2903735339641571,
885
+ 0.3420460522174835,
886
+ 0.3008522093296051,
887
+ 0.33471688628196716,
888
+ 0.3294910788536072,
889
+ 0.32927003502845764,
890
+ 0.32471102476119995,
891
+ 0.28034111857414246,
892
+ 0.3116665184497833,
893
+ 0.2901412546634674,
894
+ 0.28904274106025696,
895
+ 0.32102224230766296,
896
+ 0.30633339285850525,
897
+ 0.3276800513267517,
898
+ 0.2971253991127014,
899
+ 0.2962017059326172,
900
+ 0.3231637179851532,
901
+ 0.32558923959732056,
902
+ 0.3026246726512909,
903
+ 0.29656192660331726,
904
+ 0.2870340645313263,
905
+ 0.3136325478553772,
906
+ 0.3107961118221283,
907
+ 0.27719056606292725,
908
+ 0.28867000341415405,
909
+ 0.3126758933067322,
910
+ 0.3061710000038147,
911
+ 0.2899143397808075,
912
+ 0.29927772283554077,
913
+ 0.33476075530052185,
914
+ 0.303933709859848,
915
+ 0.30387571454048157,
916
+ 0.27551040053367615,
917
+ 0.3283371329307556,
918
+ 0.28035324811935425,
919
+ 0.27137964963912964,
920
+ 0.2864993214607239,
921
+ 0.28508108854293823,
922
+ 0.28911978006362915,
923
+ 0.3033185601234436,
924
+ 0.33768197894096375,
925
+ 0.2865414619445801,
926
+ 0.27776697278022766,
927
+ 0.2693435847759247,
928
+ 0.2960215210914612,
929
+ 0.30107125639915466,
930
+ 0.3116464912891388,
931
+ 0.31165823340415955,
932
+ 0.27068740129470825,
933
+ 0.34802037477493286,
934
+ 0.26017308235168457,
935
+ 0.2988985478878021,
936
+ 0.29535359144210815,
937
+ 0.2850148379802704,
938
+ 0.28126442432403564,
939
+ 0.24892422556877136,
940
+ 0.27116042375564575,
941
+ 0.29559674859046936,
942
+ 0.2617887854576111,
943
+ 0.2912878096103668,
944
+ 0.28404542803764343,
945
+ 0.2802587151527405,
946
+ 0.29184287786483765,
947
+ 0.25255274772644043,
948
+ 0.28309065103530884,
949
+ 0.2794201374053955,
950
+ 0.32000473141670227,
951
+ 0.2768300175666809,
952
+ 0.2603008449077606,
953
+ 0.28732627630233765,
954
+ 0.2696552872657776,
955
+ 0.28202253580093384,
956
+ 0.286325067281723,
957
+ 0.2886154353618622,
958
+ 0.2635592520236969,
959
+ 0.25940585136413574,
960
+ 0.3054310381412506,
961
+ 0.2695368230342865,
962
+ 0.2958202660083771,
963
+ 0.275333046913147,
964
+ 0.2526370882987976,
965
+ 0.2347157746553421,
966
+ 0.3015449345111847,
967
+ 0.26805999875068665,
968
+ 0.252463698387146,
969
+ 0.2655087411403656,
970
+ 0.26605311036109924,
971
+ 0.2699032425880432,
972
+ 0.2725366950035095,
973
+ 0.24781785905361176,
974
+ 0.2585459053516388,
975
+ 0.2588556706905365,
976
+ 0.2329813539981842,
977
+ 0.2423897236585617,
978
+ 0.2820303440093994,
979
+ 0.2728225588798523,
980
+ 0.2572624981403351,
981
+ 0.24342893064022064,
982
+ 0.2748130261898041,
983
+ 0.2646864950656891,
984
+ 0.2545367181301117,
985
+ 0.26127490401268005,
986
+ 0.2617938220500946,
987
+ 0.2913815379142761,
988
+ 0.2748410403728485,
989
+ 0.2587568461894989,
990
+ 0.29032665491104126,
991
+ 0.26374754309654236,
992
+ 0.3073883652687073,
993
+ 0.24755597114562988,
994
+ 0.24816052615642548,
995
+ 0.30302050709724426,
996
+ 0.2414945363998413,
997
+ 0.23515714704990387,
998
+ 0.24183964729309082,
999
+ 0.272430956363678,
1000
+ 0.2524668872356415,
1001
+ 0.2560848593711853,
1002
+ 0.264630526304245,
1003
+ 0.26772230863571167,
1004
+ 0.25159573554992676,
1005
+ 0.2683458626270294,
1006
+ 0.2672760486602783,
1007
+ 0.26370224356651306,
1008
+ 0.2499549686908722,
1009
+ 0.23869001865386963,
1010
+ 0.26777559518814087,
1011
+ 0.26570749282836914,
1012
+ 0.2595261037349701,
1013
+ 0.24541568756103516,
1014
+ 0.2415076494216919,
1015
+ 0.26374584436416626,
1016
+ 0.2645344138145447,
1017
+ 0.2518458664417267,
1018
+ 0.25503847002983093,
1019
+ 0.2291836440563202,
1020
+ 0.23590432107448578,
1021
+ 0.2632165849208832,
1022
+ 0.2532927095890045,
1023
+ 0.2827414572238922,
1024
+ 0.25054121017456055,
1025
+ 0.23753012716770172,
1026
+ 0.25390326976776123,
1027
+ 0.22770343720912933,
1028
+ 0.24350132048130035,
1029
+ 0.2328876405954361,
1030
+ 0.2645018696784973,
1031
+ 0.2541333734989166,
1032
+ 0.23772020637989044,
1033
+ 0.2970789968967438,
1034
+ 0.2636003792285919,
1035
+ 0.2570382058620453,
1036
+ 0.25797900557518005,
1037
+ 0.23588991165161133,
1038
+ 0.23865680396556854,
1039
+ 0.24991001188755035,
1040
+ 0.2434508055448532,
1041
+ 0.27677208185195923,
1042
+ 0.23119306564331055,
1043
+ 0.25682756304740906,
1044
+ 0.24296820163726807,
1045
+ 0.27047213912010193,
1046
+ 0.27801012992858887,
1047
+ 0.264803022146225,
1048
+ 0.2683183550834656,
1049
+ 0.26446881890296936,
1050
+ 0.25798138976097107,
1051
+ 0.25945737957954407,
1052
+ 0.23650023341178894,
1053
+ 0.23238001763820648,
1054
+ 0.2904701232910156,
1055
+ 0.2603924572467804,
1056
+ 0.26522642374038696,
1057
+ 0.22868548333644867,
1058
+ 0.23852208256721497,
1059
+ 0.2342144101858139,
1060
+ 0.23408107459545135,
1061
+ 0.23311002552509308,
1062
+ 0.23117592930793762,
1063
+ 0.21147707104682922,
1064
+ 0.2483009248971939,
1065
+ 0.22474507987499237,
1066
+ 0.25965139269828796,
1067
+ 0.2393711805343628,
1068
+ 0.2501153349876404,
1069
+ 0.2677519619464874,
1070
+ 0.2438044548034668,
1071
+ 0.2594291865825653,
1072
+ 0.22897210717201233,
1073
+ 0.2529992163181305,
1074
+ 0.23063622415065765,
1075
+ 0.2547101378440857,
1076
+ 0.2712981700897217,
1077
+ 0.24938587844371796,
1078
+ 0.21392764151096344,
1079
+ 0.25845345854759216,
1080
+ 0.23748472332954407,
1081
+ 0.22717954218387604,
1082
+ 0.265554815530777,
1083
+ 0.2474057674407959,
1084
+ 0.25448599457740784,
1085
+ 0.23507292568683624,
1086
+ 0.24866999685764313,
1087
+ 0.2171236276626587,
1088
+ 0.2446296215057373,
1089
+ 0.22870604693889618,
1090
+ 0.23484086990356445,
1091
+ 0.1889449656009674,
1092
+ 0.24022945761680603,
1093
+ 0.23945651948451996,
1094
+ 0.24334020912647247,
1095
+ 0.21448609232902527,
1096
+ 0.23696620762348175,
1097
+ 0.2275967299938202,
1098
+ 0.24452278017997742,
1099
+ 0.2396339625120163,
1100
+ 0.22207772731781006,
1101
+ 0.2181873768568039,
1102
+ 0.26768797636032104,
1103
+ 0.2233200967311859,
1104
+ 0.24209947884082794,
1105
+ 0.22863610088825226,
1106
+ 0.25096267461776733,
1107
+ 0.2189108282327652,
1108
+ 0.26266658306121826,
1109
+ 0.23849013447761536,
1110
+ 0.25674381852149963,
1111
+ 0.2303982824087143,
1112
+ 0.20671793818473816,
1113
+ 0.22474977374076843,
1114
+ 0.2159137725830078,
1115
+ 0.229452446103096,
1116
+ 0.2287021428346634,
1117
+ 0.24896666407585144,
1118
+ 0.23080256581306458,
1119
+ 0.23838010430335999,
1120
+ 0.24007712304592133,
1121
+ 0.25502443313598633,
1122
+ 0.2228824347257614,
1123
+ 0.23436620831489563,
1124
+ 0.2376219928264618,
1125
+ 0.2293662279844284,
1126
+ 0.2532448470592499,
1127
+ 0.23155486583709717,
1128
+ 0.2602384388446808,
1129
+ 0.24036096036434174,
1130
+ 0.24773462116718292,
1131
+ 0.26077893376350403,
1132
+ 0.21069155633449554,
1133
+ 0.2436290979385376,
1134
+ 0.21279123425483704,
1135
+ 0.23200498521327972,
1136
+ 0.2329494208097458,
1137
+ 0.24477998912334442,
1138
+ 0.25203460454940796,
1139
+ 0.2424793690443039,
1140
+ 0.23813113570213318,
1141
+ 0.24559161067008972,
1142
+ 0.2250899076461792,
1143
+ 0.2363835871219635,
1144
+ 0.2444305121898651,
1145
+ 0.24078068137168884,
1146
+ 0.2577361464500427,
1147
+ 0.2392052859067917,
1148
+ 0.2356582134962082,
1149
+ 0.23357577621936798,
1150
+ 0.23279152810573578,
1151
+ 0.21853098273277283,
1152
+ 0.23393948376178741,
1153
+ 0.23393060266971588,
1154
+ 0.2311018854379654,
1155
+ 0.22463004291057587,
1156
+ 0.24547739326953888,
1157
+ 0.21430929005146027,
1158
+ 0.23214514553546906,
1159
+ 0.2253446877002716,
1160
+ 0.2570059299468994,
1161
+ 0.2428572028875351,
1162
+ 0.23149731755256653,
1163
+ 0.2622249722480774,
1164
+ 0.22553865611553192,
1165
+ 0.24374201893806458,
1166
+ 0.24300014972686768,
1167
+ 0.21938133239746094,
1168
+ 0.21731063723564148,
1169
+ 0.2513446807861328,
1170
+ 0.23932620882987976,
1171
+ 0.2427012175321579,
1172
+ 0.23802350461483002,
1173
+ 0.24245336651802063,
1174
+ 0.21608999371528625,
1175
+ 0.22886040806770325,
1176
+ 0.2449028193950653,
1177
+ 0.21934141218662262,
1178
+ 0.23627139627933502,
1179
+ 0.20033489167690277,
1180
+ 0.23405806720256805,
1181
+ 0.23975898325443268,
1182
+ 0.24200403690338135,
1183
+ 0.23389966785907745,
1184
+ 0.24207845330238342,
1185
+ 0.24395427107810974,
1186
+ 0.24277034401893616,
1187
+ 0.2291681468486786,
1188
+ 0.2559303343296051,
1189
+ 0.21439337730407715,
1190
+ 0.23189301788806915,
1191
+ 0.24813392758369446,
1192
+ 0.2403029501438141,
1193
+ 0.21944734454154968,
1194
+ 0.26975420117378235,
1195
+ 0.2539968192577362,
1196
+ 0.2568897604942322,
1197
+ 0.25361186265945435,
1198
+ 0.2449120432138443,
1199
+ 0.23540370166301727,
1200
+ 0.2224510759115219,
1201
+ 0.23179641366004944,
1202
+ 0.20641081035137177,
1203
+ 0.22924134135246277,
1204
+ 0.219978928565979,
1205
+ 0.2511546015739441,
1206
+ 0.2110583484172821,
1207
+ 0.22338168323040009,
1208
+ 0.25245752930641174,
1209
+ 0.22505000233650208,
1210
+ 0.2245175987482071,
1211
+ 0.23512394726276398,
1212
+ 0.21569010615348816,
1213
+ 0.24025654792785645,
1214
+ 0.2134380042552948,
1215
+ 0.19874753057956696,
1216
+ 0.20266090333461761,
1217
+ 0.22143004834651947,
1218
+ 0.22060725092887878,
1219
+ 0.22482965886592865,
1220
+ 0.20826180279254913,
1221
+ 0.2163400650024414,
1222
+ 0.2172548472881317,
1223
+ 0.2323136329650879,
1224
+ 0.23619842529296875,
1225
+ 0.2161109298467636,
1226
+ 0.23117096722126007,
1227
+ 0.22682513296604156,
1228
+ 0.23544420301914215,
1229
+ 0.20841632783412933,
1230
+ 0.23073826730251312,
1231
+ 0.23295699059963226,
1232
+ 0.24892571568489075,
1233
+ 0.217569038271904,
1234
+ 0.20324870944023132,
1235
+ 0.22242750227451324,
1236
+ 0.21749792993068695,
1237
+ 0.21960806846618652,
1238
+ 0.23340265452861786,
1239
+ 0.21995973587036133,
1240
+ 0.2608080506324768,
1241
+ 0.21185769140720367,
1242
+ 0.2387785017490387,
1243
+ 0.2483835518360138,
1244
+ 0.19331252574920654,
1245
+ 0.2512570321559906,
1246
+ 0.22524642944335938,
1247
+ 0.22987328469753265,
1248
+ 0.23031756281852722,
1249
+ 0.271034300327301,
1250
+ 0.25407272577285767,
1251
+ 0.23441484570503235,
1252
+ 0.19989612698554993,
1253
+ 0.24161195755004883,
1254
+ 0.22557173669338226,
1255
+ 0.25727805495262146
1256
  ],
1257
  "base_loss": [
1258
+ 11.963690757751465,
1259
+ 11.661938667297363,
1260
+ 11.223055839538574,
1261
+ 10.688573837280273,
1262
+ 9.94072437286377,
1263
+ 9.775545120239258,
1264
+ 9.39625072479248,
1265
+ 9.00570011138916,
1266
+ 8.761700630187988,
1267
+ 8.4192533493042,
1268
+ 8.12822437286377,
1269
+ 7.931671142578125,
1270
+ 7.313607215881348,
1271
+ 6.978346347808838,
1272
+ 6.648111820220947,
1273
+ 5.993403911590576,
1274
+ 5.525230407714844,
1275
+ 5.094429969787598,
1276
+ 4.402842998504639,
1277
+ 3.949738025665283,
1278
+ 3.4472267627716064,
1279
+ 2.9571051597595215,
1280
+ 2.724224328994751,
1281
+ 2.427931070327759,
1282
+ 2.2689688205718994,
1283
+ 2.1323626041412354,
1284
+ 2.0063560009002686,
1285
+ 2.0175435543060303,
1286
+ 1.9365376234054565,
1287
+ 1.9525864124298096,
1288
+ 1.9771562814712524,
1289
+ 1.9292795658111572,
1290
+ 1.9087927341461182,
1291
+ 1.8182607889175415,
1292
+ 1.8655515909194946,
1293
+ 1.7887706756591797,
1294
+ 1.8116521835327148,
1295
+ 1.8060662746429443,
1296
+ 1.8420522212982178,
1297
+ 1.8362396955490112,
1298
+ 1.7543896436691284,
1299
+ 1.7854712009429932,
1300
+ 1.838714599609375,
1301
+ 1.8220285177230835,
1302
+ 1.824865460395813,
1303
+ 1.744118332862854,
1304
+ 1.7437489032745361,
1305
+ 1.8406013250350952,
1306
+ 1.769796371459961,
1307
+ 1.7702674865722656,
1308
+ 1.6466058492660522,
1309
+ 1.8461577892303467,
1310
+ 1.7179840803146362,
1311
+ 1.6748616695404053,
1312
+ 1.6761716604232788,
1313
+ 1.7217031717300415,
1314
+ 1.7508471012115479,
1315
+ 1.7157037258148193,
1316
+ 1.6086606979370117,
1317
+ 1.6273046731948853,
1318
+ 1.6577644348144531,
1319
+ 1.5764987468719482,
1320
+ 1.6589974164962769,
1321
+ 1.616567850112915,
1322
+ 1.6576578617095947,
1323
+ 1.4999885559082031,
1324
+ 1.5213489532470703,
1325
+ 1.5925500392913818,
1326
+ 1.4423539638519287,
1327
+ 1.528339147567749,
1328
+ 1.402854323387146,
1329
+ 1.3988912105560303,
1330
+ 1.367125153541565,
1331
+ 1.2756946086883545,
1332
+ 1.2506844997406006,
1333
+ 1.173041582107544,
1334
+ 1.193954586982727,
1335
+ 0.990950882434845,
1336
+ 0.9718019366264343,
1337
+ 0.943374752998352,
1338
+ 0.943237841129303,
1339
+ 0.8248871564865112,
1340
+ 0.9327092170715332,
1341
+ 0.7993831634521484,
1342
+ 0.7706826329231262,
1343
+ 0.7637476325035095,
1344
+ 0.7574962973594666,
1345
+ 0.7494969367980957,
1346
+ 0.7162280678749084,
1347
+ 0.7451773881912231,
1348
+ 0.670974612236023,
1349
+ 0.663203239440918,
1350
+ 0.6701799035072327,
1351
+ 0.6127704381942749,
1352
+ 0.6420816779136658,
1353
+ 0.6314144134521484,
1354
+ 0.6287131905555725,
1355
+ 0.5879684090614319,
1356
+ 0.6163527369499207,
1357
+ 0.598551869392395,
1358
+ 0.6191786527633667,
1359
+ 0.5709514617919922,
1360
+ 0.5744446516036987,
1361
+ 0.5953712463378906,
1362
+ 0.5669698119163513,
1363
+ 0.569784939289093,
1364
+ 0.590087354183197,
1365
+ 0.6043781042098999,
1366
+ 0.5754431486129761,
1367
+ 0.5330023169517517,
1368
+ 0.5477746725082397,
1369
+ 0.5358116626739502,
1370
+ 0.5499526262283325,
1371
+ 0.5191396474838257,
1372
+ 0.5597048997879028,
1373
+ 0.5326811671257019,
1374
+ 0.5370150804519653,
1375
+ 0.547336995601654,
1376
+ 0.5351187586784363,
1377
+ 0.5349653959274292,
1378
+ 0.5326358079910278,
1379
+ 0.5140055418014526,
1380
+ 0.5046959519386292,
1381
+ 0.49848243594169617,
1382
+ 0.5145978927612305,
1383
+ 0.489215224981308,
1384
+ 0.5175383687019348,
1385
+ 0.48192352056503296,
1386
+ 0.5176753401756287,
1387
+ 0.4990863502025604,
1388
+ 0.49212411046028137,
1389
+ 0.5086663365364075,
1390
+ 0.4860772490501404,
1391
+ 0.5174476504325867,
1392
+ 0.5017643570899963,
1393
+ 0.4740331470966339,
1394
+ 0.46239525079727173,
1395
+ 0.4612799286842346,
1396
+ 0.4830116629600525,
1397
+ 0.464834064245224,
1398
+ 0.4621468186378479,
1399
+ 0.46891137957572937,
1400
+ 0.5075568556785583,
1401
+ 0.43174198269844055,
1402
+ 0.46206027269363403,
1403
+ 0.4287765920162201,
1404
+ 0.4551263451576233,
1405
+ 0.44393548369407654,
1406
+ 0.43048691749572754,
1407
+ 0.4558803141117096,
1408
+ 0.4532429873943329,
1409
+ 0.4646574556827545,
1410
+ 0.431844025850296,
1411
+ 0.4672628939151764,
1412
+ 0.44547510147094727,
1413
+ 0.4079582691192627,
1414
+ 0.4426649212837219,
1415
+ 0.39009925723075867,
1416
+ 0.43490490317344666,
1417
+ 0.44311219453811646,
1418
+ 0.39784926176071167,
1419
+ 0.41195815801620483,
1420
+ 0.42954376339912415,
1421
+ 0.41115802526474,
1422
+ 0.4012797176837921,
1423
+ 0.38349026441574097,
1424
+ 0.43120235204696655,
1425
+ 0.39938807487487793,
1426
+ 0.4433351457118988,
1427
+ 0.4032042324542999,
1428
+ 0.4240035116672516,
1429
+ 0.4189927577972412,
1430
+ 0.3758549094200134,
1431
+ 0.4081072211265564,
1432
+ 0.41931289434432983,
1433
+ 0.396896094083786,
1434
+ 0.37466028332710266,
1435
+ 0.41262954473495483,
1436
+ 0.3888600766658783,
1437
+ 0.38536521792411804,
1438
+ 0.4092770516872406,
1439
+ 0.3643164336681366,
1440
+ 0.3778727054595947,
1441
+ 0.3487100303173065,
1442
+ 0.38028067350387573,
1443
+ 0.36624351143836975,
1444
+ 0.3879108428955078,
1445
+ 0.3812934458255768,
1446
+ 0.4147397577762604,
1447
+ 0.32763606309890747,
1448
+ 0.362922728061676,
1449
+ 0.37037578225135803,
1450
+ 0.3684242069721222,
1451
+ 0.3541705906391144,
1452
+ 0.3531418442726135,
1453
+ 0.38803714513778687,
1454
+ 0.3663487434387207,
1455
+ 0.34989604353904724,
1456
+ 0.3375271260738373,
1457
+ 0.3937970995903015,
1458
+ 0.36703363060951233,
1459
+ 0.36325493454933167,
1460
+ 0.35152706503868103,
1461
+ 0.36929401755332947,
1462
+ 0.3409608006477356,
1463
+ 0.3745754659175873,
1464
+ 0.36702921986579895,
1465
+ 0.36814290285110474,
1466
+ 0.34607431292533875,
1467
+ 0.3489746153354645,
1468
+ 0.3475476801395416,
1469
+ 0.3784462809562683,
1470
+ 0.37143474817276,
1471
+ 0.3513146936893463,
1472
+ 0.3392271101474762,
1473
+ 0.37690967321395874,
1474
+ 0.35800108313560486,
1475
+ 0.35400205850601196,
1476
+ 0.3127591609954834,
1477
+ 0.3481731116771698,
1478
+ 0.3677438199520111,
1479
+ 0.37050268054008484,
1480
+ 0.3325396180152893,
1481
+ 0.33604928851127625,
1482
+ 0.34623295068740845,
1483
+ 0.32801350951194763,
1484
+ 0.3004973828792572,
1485
+ 0.33102330565452576,
1486
+ 0.35216864943504333,
1487
+ 0.35869526863098145,
1488
+ 0.3230397701263428,
1489
+ 0.34125009179115295,
1490
+ 0.3288026750087738,
1491
+ 0.36230114102363586,
1492
+ 0.32093361020088196,
1493
+ 0.33653420209884644,
1494
+ 0.31276172399520874,
1495
+ 0.32386818528175354,
1496
+ 0.30667203664779663,
1497
+ 0.3109426200389862,
1498
+ 0.3534684181213379,
1499
+ 0.3197288513183594,
1500
+ 0.30439478158950806,
1501
+ 0.292362242937088,
1502
+ 0.3334762752056122,
1503
+ 0.35194072127342224,
1504
+ 0.29283004999160767,
1505
+ 0.32437053322792053,
1506
+ 0.293384850025177,
1507
+ 0.31290170550346375,
1508
+ 0.32321569323539734,
1509
+ 0.32250460982322693,
1510
+ 0.3221062123775482,
1511
+ 0.2903735339641571,
1512
+ 0.3420460522174835,
1513
+ 0.3008522093296051,
1514
+ 0.33471688628196716,
1515
+ 0.3294910788536072,
1516
+ 0.32927003502845764,
1517
+ 0.32471102476119995,
1518
+ 0.28034111857414246,
1519
+ 0.3116665184497833,
1520
+ 0.2901412546634674,
1521
+ 0.28904274106025696,
1522
+ 0.32102224230766296,
1523
+ 0.30633339285850525,
1524
+ 0.3276800513267517,
1525
+ 0.2971253991127014,
1526
+ 0.2962017059326172,
1527
+ 0.3231637179851532,
1528
+ 0.32558923959732056,
1529
+ 0.3026246726512909,
1530
+ 0.29656192660331726,
1531
+ 0.2870340645313263,
1532
+ 0.3136325478553772,
1533
+ 0.3107961118221283,
1534
+ 0.27719056606292725,
1535
+ 0.28867000341415405,
1536
+ 0.3126758933067322,
1537
+ 0.3061710000038147,
1538
+ 0.2899143397808075,
1539
+ 0.29927772283554077,
1540
+ 0.33476075530052185,
1541
+ 0.303933709859848,
1542
+ 0.30387571454048157,
1543
+ 0.27551040053367615,
1544
+ 0.3283371329307556,
1545
+ 0.28035324811935425,
1546
+ 0.27137964963912964,
1547
+ 0.2864993214607239,
1548
+ 0.28508108854293823,
1549
+ 0.28911978006362915,
1550
+ 0.3033185601234436,
1551
+ 0.33768197894096375,
1552
+ 0.2865414619445801,
1553
+ 0.27776697278022766,
1554
+ 0.2693435847759247,
1555
+ 0.2960215210914612,
1556
+ 0.30107125639915466,
1557
+ 0.3116464912891388,
1558
+ 0.31165823340415955,
1559
+ 0.27068740129470825,
1560
+ 0.34802037477493286,
1561
+ 0.26017308235168457,
1562
+ 0.2988985478878021,
1563
+ 0.29535359144210815,
1564
+ 0.2850148379802704,
1565
+ 0.28126442432403564,
1566
+ 0.24892422556877136,
1567
+ 0.27116042375564575,
1568
+ 0.29559674859046936,
1569
+ 0.2617887854576111,
1570
+ 0.2912878096103668,
1571
+ 0.28404542803764343,
1572
+ 0.2802587151527405,
1573
+ 0.29184287786483765,
1574
+ 0.25255274772644043,
1575
+ 0.28309065103530884,
1576
+ 0.2794201374053955,
1577
+ 0.32000473141670227,
1578
+ 0.2768300175666809,
1579
+ 0.2603008449077606,
1580
+ 0.28732627630233765,
1581
+ 0.2696552872657776,
1582
+ 0.28202253580093384,
1583
+ 0.286325067281723,
1584
+ 0.2886154353618622,
1585
+ 0.2635592520236969,
1586
+ 0.25940585136413574,
1587
+ 0.3054310381412506,
1588
+ 0.2695368230342865,
1589
+ 0.2958202660083771,
1590
+ 0.275333046913147,
1591
+ 0.2526370882987976,
1592
+ 0.2347157746553421,
1593
+ 0.3015449345111847,
1594
+ 0.26805999875068665,
1595
+ 0.252463698387146,
1596
+ 0.2655087411403656,
1597
+ 0.26605311036109924,
1598
+ 0.2699032425880432,
1599
+ 0.2725366950035095,
1600
+ 0.24781785905361176,
1601
+ 0.2585459053516388,
1602
+ 0.2588556706905365,
1603
+ 0.2329813539981842,
1604
+ 0.2423897236585617,
1605
+ 0.2820303440093994,
1606
+ 0.2728225588798523,
1607
+ 0.2572624981403351,
1608
+ 0.24342893064022064,
1609
+ 0.2748130261898041,
1610
+ 0.2646864950656891,
1611
+ 0.2545367181301117,
1612
+ 0.26127490401268005,
1613
+ 0.2617938220500946,
1614
+ 0.2913815379142761,
1615
+ 0.2748410403728485,
1616
+ 0.2587568461894989,
1617
+ 0.29032665491104126,
1618
+ 0.26374754309654236,
1619
+ 0.3073883652687073,
1620
+ 0.24755597114562988,
1621
+ 0.24816052615642548,
1622
+ 0.30302050709724426,
1623
+ 0.2414945363998413,
1624
+ 0.23515714704990387,
1625
+ 0.24183964729309082,
1626
+ 0.272430956363678,
1627
+ 0.2524668872356415,
1628
+ 0.2560848593711853,
1629
+ 0.264630526304245,
1630
+ 0.26772230863571167,
1631
+ 0.25159573554992676,
1632
+ 0.2683458626270294,
1633
+ 0.2672760486602783,
1634
+ 0.26370224356651306,
1635
+ 0.2499549686908722,
1636
+ 0.23869001865386963,
1637
+ 0.26777559518814087,
1638
+ 0.26570749282836914,
1639
+ 0.2595261037349701,
1640
+ 0.24541568756103516,
1641
+ 0.2415076494216919,
1642
+ 0.26374584436416626,
1643
+ 0.2645344138145447,
1644
+ 0.2518458664417267,
1645
+ 0.25503847002983093,
1646
+ 0.2291836440563202,
1647
+ 0.23590432107448578,
1648
+ 0.2632165849208832,
1649
+ 0.2532927095890045,
1650
+ 0.2827414572238922,
1651
+ 0.25054121017456055,
1652
+ 0.23753012716770172,
1653
+ 0.25390326976776123,
1654
+ 0.22770343720912933,
1655
+ 0.24350132048130035,
1656
+ 0.2328876405954361,
1657
+ 0.2645018696784973,
1658
+ 0.2541333734989166,
1659
+ 0.23772020637989044,
1660
+ 0.2970789968967438,
1661
+ 0.2636003792285919,
1662
+ 0.2570382058620453,
1663
+ 0.25797900557518005,
1664
+ 0.23588991165161133,
1665
+ 0.23865680396556854,
1666
+ 0.24991001188755035,
1667
+ 0.2434508055448532,
1668
+ 0.27677208185195923,
1669
+ 0.23119306564331055,
1670
+ 0.25682756304740906,
1671
+ 0.24296820163726807,
1672
+ 0.27047213912010193,
1673
+ 0.27801012992858887,
1674
+ 0.264803022146225,
1675
+ 0.2683183550834656,
1676
+ 0.26446881890296936,
1677
+ 0.25798138976097107,
1678
+ 0.25945737957954407,
1679
+ 0.23650023341178894,
1680
+ 0.23238001763820648,
1681
+ 0.2904701232910156,
1682
+ 0.2603924572467804,
1683
+ 0.26522642374038696,
1684
+ 0.22868548333644867,
1685
+ 0.23852208256721497,
1686
+ 0.2342144101858139,
1687
+ 0.23408107459545135,
1688
+ 0.23311002552509308,
1689
+ 0.23117592930793762,
1690
+ 0.21147707104682922,
1691
+ 0.2483009248971939,
1692
+ 0.22474507987499237,
1693
+ 0.25965139269828796,
1694
+ 0.2393711805343628,
1695
+ 0.2501153349876404,
1696
+ 0.2677519619464874,
1697
+ 0.2438044548034668,
1698
+ 0.2594291865825653,
1699
+ 0.22897210717201233,
1700
+ 0.2529992163181305,
1701
+ 0.23063622415065765,
1702
+ 0.2547101378440857,
1703
+ 0.2712981700897217,
1704
+ 0.24938587844371796,
1705
+ 0.21392764151096344,
1706
+ 0.25845345854759216,
1707
+ 0.23748472332954407,
1708
+ 0.22717954218387604,
1709
+ 0.265554815530777,
1710
+ 0.2474057674407959,
1711
+ 0.25448599457740784,
1712
+ 0.23507292568683624,
1713
+ 0.24866999685764313,
1714
+ 0.2171236276626587,
1715
+ 0.2446296215057373,
1716
+ 0.22870604693889618,
1717
+ 0.23484086990356445,
1718
+ 0.1889449656009674,
1719
+ 0.24022945761680603,
1720
+ 0.23945651948451996,
1721
+ 0.24334020912647247,
1722
+ 0.21448609232902527,
1723
+ 0.23696620762348175,
1724
+ 0.2275967299938202,
1725
+ 0.24452278017997742,
1726
+ 0.2396339625120163,
1727
+ 0.22207772731781006,
1728
+ 0.2181873768568039,
1729
+ 0.26768797636032104,
1730
+ 0.2233200967311859,
1731
+ 0.24209947884082794,
1732
+ 0.22863610088825226,
1733
+ 0.25096267461776733,
1734
+ 0.2189108282327652,
1735
+ 0.26266658306121826,
1736
+ 0.23849013447761536,
1737
+ 0.25674381852149963,
1738
+ 0.2303982824087143,
1739
+ 0.20671793818473816,
1740
+ 0.22474977374076843,
1741
+ 0.2159137725830078,
1742
+ 0.229452446103096,
1743
+ 0.2287021428346634,
1744
+ 0.24896666407585144,
1745
+ 0.23080256581306458,
1746
+ 0.23838010430335999,
1747
+ 0.24007712304592133,
1748
+ 0.25502443313598633,
1749
+ 0.2228824347257614,
1750
+ 0.23436620831489563,
1751
+ 0.2376219928264618,
1752
+ 0.2293662279844284,
1753
+ 0.2532448470592499,
1754
+ 0.23155486583709717,
1755
+ 0.2602384388446808,
1756
+ 0.24036096036434174,
1757
+ 0.24773462116718292,
1758
+ 0.26077893376350403,
1759
+ 0.21069155633449554,
1760
+ 0.2436290979385376,
1761
+ 0.21279123425483704,
1762
+ 0.23200498521327972,
1763
+ 0.2329494208097458,
1764
+ 0.24477998912334442,
1765
+ 0.25203460454940796,
1766
+ 0.2424793690443039,
1767
+ 0.23813113570213318,
1768
+ 0.24559161067008972,
1769
+ 0.2250899076461792,
1770
+ 0.2363835871219635,
1771
+ 0.2444305121898651,
1772
+ 0.24078068137168884,
1773
+ 0.2577361464500427,
1774
+ 0.2392052859067917,
1775
+ 0.2356582134962082,
1776
+ 0.23357577621936798,
1777
+ 0.23279152810573578,
1778
+ 0.21853098273277283,
1779
+ 0.23393948376178741,
1780
+ 0.23393060266971588,
1781
+ 0.2311018854379654,
1782
+ 0.22463004291057587,
1783
+ 0.24547739326953888,
1784
+ 0.21430929005146027,
1785
+ 0.23214514553546906,
1786
+ 0.2253446877002716,
1787
+ 0.2570059299468994,
1788
+ 0.2428572028875351,
1789
+ 0.23149731755256653,
1790
+ 0.2622249722480774,
1791
+ 0.22553865611553192,
1792
+ 0.24374201893806458,
1793
+ 0.24300014972686768,
1794
+ 0.21938133239746094,
1795
+ 0.21731063723564148,
1796
+ 0.2513446807861328,
1797
+ 0.23932620882987976,
1798
+ 0.2427012175321579,
1799
+ 0.23802350461483002,
1800
+ 0.24245336651802063,
1801
+ 0.21608999371528625,
1802
+ 0.22886040806770325,
1803
+ 0.2449028193950653,
1804
+ 0.21934141218662262,
1805
+ 0.23627139627933502,
1806
+ 0.20033489167690277,
1807
+ 0.23405806720256805,
1808
+ 0.23975898325443268,
1809
+ 0.24200403690338135,
1810
+ 0.23389966785907745,
1811
+ 0.24207845330238342,
1812
+ 0.24395427107810974,
1813
+ 0.24277034401893616,
1814
+ 0.2291681468486786,
1815
+ 0.2559303343296051,
1816
+ 0.21439337730407715,
1817
+ 0.23189301788806915,
1818
+ 0.24813392758369446,
1819
+ 0.2403029501438141,
1820
+ 0.21944734454154968,
1821
+ 0.26975420117378235,
1822
+ 0.2539968192577362,
1823
+ 0.2568897604942322,
1824
+ 0.25361186265945435,
1825
+ 0.2449120432138443,
1826
+ 0.23540370166301727,
1827
+ 0.2224510759115219,
1828
+ 0.23179641366004944,
1829
+ 0.20641081035137177,
1830
+ 0.22924134135246277,
1831
+ 0.219978928565979,
1832
+ 0.2511546015739441,
1833
+ 0.2110583484172821,
1834
+ 0.22338168323040009,
1835
+ 0.25245752930641174,
1836
+ 0.22505000233650208,
1837
+ 0.2245175987482071,
1838
+ 0.23512394726276398,
1839
+ 0.21569010615348816,
1840
+ 0.24025654792785645,
1841
+ 0.2134380042552948,
1842
+ 0.19874753057956696,
1843
+ 0.20266090333461761,
1844
+ 0.22143004834651947,
1845
+ 0.22060725092887878,
1846
+ 0.22482965886592865,
1847
+ 0.20826180279254913,
1848
+ 0.2163400650024414,
1849
+ 0.2172548472881317,
1850
+ 0.2323136329650879,
1851
+ 0.23619842529296875,
1852
+ 0.2161109298467636,
1853
+ 0.23117096722126007,
1854
+ 0.22682513296604156,
1855
+ 0.23544420301914215,
1856
+ 0.20841632783412933,
1857
+ 0.23073826730251312,
1858
+ 0.23295699059963226,
1859
+ 0.24892571568489075,
1860
+ 0.217569038271904,
1861
+ 0.20324870944023132,
1862
+ 0.22242750227451324,
1863
+ 0.21749792993068695,
1864
+ 0.21960806846618652,
1865
+ 0.23340265452861786,
1866
+ 0.21995973587036133,
1867
+ 0.2608080506324768,
1868
+ 0.21185769140720367,
1869
+ 0.2387785017490387,
1870
+ 0.2483835518360138,
1871
+ 0.19331252574920654,
1872
+ 0.2512570321559906,
1873
+ 0.22524642944335938,
1874
+ 0.22987328469753265,
1875
+ 0.23031756281852722,
1876
+ 0.271034300327301,
1877
+ 0.25407272577285767,
1878
+ 0.23441484570503235,
1879
+ 0.19989612698554993,
1880
+ 0.24161195755004883,
1881
+ 0.22557173669338226,
1882
+ 0.25727805495262146
1883
  ],
1884
  "lr": [
1885
  1.0458911419423694e-06,
 
2553
  20
2554
  ],
2555
  "eval_accuracy": [
2556
+ 0.003157894736842105,
2557
+ 0.011578947368421053,
2558
+ 0.056842105263157895,
2559
+ 0.11473684210526315,
2560
+ 0.15052631578947367,
2561
+ 0.21894736842105264,
2562
+ 0.3063157894736842,
2563
+ 0.3136842105263158,
2564
+ 0.35368421052631577,
2565
+ 0.36210526315789476,
2566
+ 0.32526315789473687,
2567
+ 0.39473684210526316,
2568
+ 0.41789473684210526,
2569
+ 0.4368421052631579,
2570
+ 0.4431578947368421,
2571
+ 0.4063157894736842,
2572
+ 0.4473684210526316,
2573
+ 0.41578947368421054,
2574
+ 0.43894736842105264,
2575
+ 0.4357894736842105
2576
  ]
2577
  },
2578
+ "final_accuracy": 0.3565384615384615,
2579
  "sft_eval": {
2580
  "config": {
2581
  "ops": "add_sub",
2582
  "K": null,
2583
  "mode": "sft",
2584
  "n_digits": 6,
2585
+ "n_per_split": 100
2586
  },
2587
  "splits": {
2588
  "add_S0": {
2589
+ "full_accuracy": 0.72,
2590
+ "digit_accuracy": 0.9571428571428572,
2591
+ "n_examples": 100,
2592
  "per_subtask": {
2593
  "SA": {
2594
+ "accuracy": 0.9537190082644628,
2595
+ "count": 605
2596
  },
2597
  "SS": {
2598
+ "accuracy": 0.9789473684210527,
2599
+ "count": 95
2600
  }
2601
  }
2602
  },
2603
  "add_S1": {
2604
+ "full_accuracy": 0.56,
2605
+ "digit_accuracy": 0.92,
2606
+ "n_examples": 100,
2607
  "per_subtask": {
2608
  "SA": {
2609
+ "accuracy": 0.9411764705882353,
2610
+ "count": 204
2611
  },
2612
  "SC": {
2613
+ "accuracy": 0.9289940828402367,
2614
+ "count": 169
2615
  },
2616
  "SS": {
2617
+ "accuracy": 0.9354838709677419,
2618
+ "count": 31
2619
  },
2620
  "UC": {
2621
+ "accuracy": 0.8986486486486487,
2622
+ "count": 296
2623
  }
2624
  }
2625
  },
2626
  "add_S2": {
2627
+ "full_accuracy": 0.31,
2628
+ "digit_accuracy": 0.8557142857142858,
2629
+ "n_examples": 100,
2630
  "per_subtask": {
2631
  "SA": {
2632
+ "accuracy": 0.9447852760736196,
2633
+ "count": 163
2634
  },
2635
  "SC": {
2636
+ "accuracy": 0.8692307692307693,
2637
+ "count": 130
2638
  },
2639
  "SS": {
2640
+ "accuracy": 0.8620689655172413,
2641
+ "count": 87
2642
  },
2643
  "UC": {
2644
+ "accuracy": 0.7192118226600985,
2645
+ "count": 203
2646
  },
2647
  "US": {
2648
+ "accuracy": 0.9487179487179487,
2649
+ "count": 117
2650
  }
2651
  }
2652
  },
2653
  "add_S3": {
2654
  "full_accuracy": 0.18,
2655
+ "digit_accuracy": 0.7514285714285714,
2656
+ "n_examples": 100,
2657
  "per_subtask": {
2658
  "SA": {
2659
+ "accuracy": 0.9669421487603306,
2660
+ "count": 121
2661
  },
2662
  "SC": {
2663
+ "accuracy": 0.9008264462809917,
2664
+ "count": 121
2665
  },
2666
  "SS": {
2667
+ "accuracy": 0.8979591836734694,
2668
+ "count": 49
2669
  },
2670
  "UC": {
2671
+ "accuracy": 0.6021505376344086,
2672
+ "count": 186
2673
  },
2674
  "US": {
2675
+ "accuracy": 0.6457399103139013,
2676
+ "count": 223
2677
  }
2678
  }
2679
  },
2680
  "add_S4": {
2681
+ "full_accuracy": 0.16,
2682
+ "digit_accuracy": 0.6514285714285715,
2683
+ "n_examples": 100,
2684
  "per_subtask": {
2685
  "SA": {
2686
+ "accuracy": 0.9807692307692307,
2687
+ "count": 104
2688
  },
2689
  "SC": {
2690
+ "accuracy": 0.9339622641509434,
2691
+ "count": 106
2692
  },
2693
  "SS": {
2694
+ "accuracy": 0.9130434782608695,
2695
+ "count": 23
2696
  },
2697
  "UC": {
2698
+ "accuracy": 0.60625,
2699
+ "count": 160
2700
  },
2701
  "US": {
2702
+ "accuracy": 0.44625407166123776,
2703
+ "count": 307
2704
  }
2705
  }
2706
  },
2707
  "add_S5": {
2708
+ "full_accuracy": 0.13,
2709
+ "digit_accuracy": 0.48142857142857143,
2710
+ "n_examples": 100,
2711
  "per_subtask": {
2712
  "SA": {
2713
  "accuracy": 1.0,
2714
+ "count": 100
2715
  },
2716
  "SC": {
2717
+ "accuracy": 0.98,
2718
+ "count": 100
2719
  },
2720
  "UC": {
2721
+ "accuracy": 0.34,
2722
+ "count": 100
2723
  },
2724
  "US": {
2725
+ "accuracy": 0.2625,
2726
+ "count": 400
2727
  }
2728
  }
2729
  },
2730
  "add_S6": {
2731
+ "full_accuracy": 0.38,
2732
+ "digit_accuracy": 0.5342857142857143,
2733
+ "n_examples": 100,
2734
  "per_subtask": {
2735
  "SC": {
2736
  "accuracy": 1.0,
2737
+ "count": 100
2738
  },
2739
  "UC": {
2740
+ "accuracy": 0.42,
2741
+ "count": 100
2742
  },
2743
  "US": {
2744
+ "accuracy": 0.464,
2745
+ "count": 500
2746
  }
2747
  }
2748
  },
2749
  "add_random": {
2750
+ "full_accuracy": 0.655,
2751
+ "digit_accuracy": 0.9435714285714286,
2752
  "n_examples": 200,
2753
  "per_subtask": {
2754
  "SA": {
2755
+ "accuracy": 0.9776286353467561,
2756
+ "count": 447
2757
  },
2758
  "SC": {
2759
+ "accuracy": 0.959375,
2760
+ "count": 320
2761
  },
2762
  "SS": {
2763
+ "accuracy": 0.9642857142857143,
2764
+ "count": 56
2765
  },
2766
  "UC": {
2767
+ "accuracy": 0.9073724007561437,
2768
+ "count": 529
2769
  },
2770
  "US": {
2771
+ "accuracy": 0.8958333333333334,
2772
+ "count": 48
2773
  }
2774
  }
2775
  },
2776
  "add_C1": {
2777
+ "full_accuracy": 0.84,
2778
+ "digit_accuracy": 0.9742857142857143,
2779
+ "n_examples": 100,
2780
  "per_subtask": {
2781
  "SA": {
2782
+ "accuracy": 0.99,
2783
+ "count": 500
2784
  },
2785
  "SC": {
2786
+ "accuracy": 0.97,
2787
+ "count": 100
2788
  },
2789
  "UC": {
2790
+ "accuracy": 0.9,
2791
+ "count": 100
2792
  }
2793
  }
2794
  },
2795
  "add_C2": {
2796
+ "full_accuracy": 0.6,
2797
+ "digit_accuracy": 0.9328571428571428,
2798
+ "n_examples": 100,
2799
  "per_subtask": {
2800
  "SA": {
2801
+ "accuracy": 0.9875,
2802
+ "count": 400
2803
  },
2804
  "SC": {
2805
+ "accuracy": 0.99,
2806
+ "count": 100
2807
  },
2808
  "UC": {
2809
+ "accuracy": 0.7756410256410257,
2810
+ "count": 156
2811
  },
2812
  "US": {
2813
+ "accuracy": 0.8636363636363636,
2814
+ "count": 44
2815
  }
2816
  }
2817
  },
2818
  "add_C3": {
2819
+ "full_accuracy": 0.33,
2820
+ "digit_accuracy": 0.8557142857142858,
2821
+ "n_examples": 100,
2822
  "per_subtask": {
2823
  "SA": {
2824
+ "accuracy": 0.9966666666666667,
2825
+ "count": 300
2826
  },
2827
  "SC": {
2828
+ "accuracy": 0.98,
2829
+ "count": 100
2830
  },
2831
  "UC": {
2832
+ "accuracy": 0.6331658291457286,
2833
+ "count": 199
2834
  },
2835
  "US": {
2836
+ "accuracy": 0.7524752475247525,
2837
+ "count": 101
2838
  }
2839
  }
2840
  },
2841
  "add_C4": {
2842
+ "full_accuracy": 0.35,
2843
+ "digit_accuracy": 0.8457142857142858,
2844
+ "n_examples": 100,
2845
  "per_subtask": {
2846
  "SA": {
2847
+ "accuracy": 1.0,
2848
+ "count": 200
2849
  },
2850
  "SC": {
2851
  "accuracy": 1.0,
2852
+ "count": 100
2853
  },
2854
  "UC": {
2855
+ "accuracy": 0.7159090909090909,
2856
+ "count": 264
2857
  },
2858
  "US": {
2859
+ "accuracy": 0.7573529411764706,
2860
+ "count": 136
2861
  }
2862
  }
2863
  },
2864
  "add_C5": {
2865
+ "full_accuracy": 0.22,
2866
+ "digit_accuracy": 0.7842857142857143,
2867
+ "n_examples": 100,
2868
  "per_subtask": {
2869
  "SA": {
2870
  "accuracy": 1.0,
2871
+ "count": 100
2872
  },
2873
  "SC": {
2874
+ "accuracy": 0.99,
2875
+ "count": 100
2876
  },
2877
  "UC": {
2878
+ "accuracy": 0.7032258064516129,
2879
+ "count": 310
2880
  },
2881
  "US": {
2882
+ "accuracy": 0.6947368421052632,
2883
+ "count": 190
2884
  }
2885
  }
2886
  },
2887
  "add_C6": {
2888
+ "full_accuracy": 0.27,
2889
+ "digit_accuracy": 0.7914285714285715,
2890
+ "n_examples": 100,
2891
  "per_subtask": {
2892
  "SC": {
2893
  "accuracy": 1.0,
2894
+ "count": 100
2895
  },
2896
  "UC": {
2897
+ "accuracy": 0.8027027027027027,
2898
+ "count": 370
2899
  },
2900
  "US": {
2901
+ "accuracy": 0.6826086956521739,
2902
+ "count": 230
2903
  }
2904
  }
2905
  },
2906
  "sub_M0": {
2907
+ "full_accuracy": 0.73,
2908
+ "digit_accuracy": 0.9557142857142857,
2909
+ "n_examples": 100,
2910
  "per_subtask": {
2911
  "MD": {
2912
+ "accuracy": 0.9512195121951219,
2913
+ "count": 615
2914
  },
2915
  "ME": {
2916
+ "accuracy": 0.9882352941176471,
2917
+ "count": 85
2918
  }
2919
  }
2920
  },
2921
  "sub_M1": {
2922
  "full_accuracy": 0.58,
2923
+ "digit_accuracy": 0.9371428571428572,
2924
+ "n_examples": 100,
2925
  "per_subtask": {
2926
  "MD": {
2927
+ "accuracy": 0.9863013698630136,
2928
+ "count": 292
2929
  },
2930
  "MB": {
2931
+ "accuracy": 0.9861111111111112,
2932
+ "count": 144
2933
  },
2934
  "ME": {
2935
  "accuracy": 1.0,
2936
+ "count": 25
2937
  },
2938
  "UB": {
2939
+ "accuracy": 0.8410041841004184,
2940
+ "count": 239
2941
  }
2942
  }
2943
  },
2944
  "sub_M2": {
2945
+ "full_accuracy": 0.17,
2946
+ "digit_accuracy": 0.8471428571428572,
2947
+ "n_examples": 100,
2948
  "per_subtask": {
2949
  "MD": {
2950
+ "accuracy": 0.981042654028436,
2951
+ "count": 211
2952
  },
2953
  "MB": {
2954
+ "accuracy": 0.9652173913043478,
2955
+ "count": 115
2956
  },
2957
  "ME": {
2958
+ "accuracy": 1.0,
2959
+ "count": 85
2960
  },
2961
  "UB": {
2962
+ "accuracy": 0.5303867403314917,
2963
+ "count": 181
2964
  },
2965
  "UD": {
2966
+ "accuracy": 0.8703703703703703,
2967
+ "count": 108
2968
  }
2969
  }
2970
  },
2971
  "sub_M3": {
2972
+ "full_accuracy": 0.02,
2973
+ "digit_accuracy": 0.6985714285714286,
2974
+ "n_examples": 100,
2975
  "per_subtask": {
2976
  "MD": {
2977
+ "accuracy": 0.9888268156424581,
2978
+ "count": 179
2979
  },
2980
  "MB": {
2981
+ "accuracy": 0.8737864077669902,
2982
+ "count": 103
2983
  },
2984
  "ME": {
2985
+ "accuracy": 1.0,
2986
+ "count": 56
2987
  },
2988
  "UB": {
2989
+ "accuracy": 0.42953020134228187,
2990
+ "count": 149
2991
  },
2992
  "UD": {
2993
+ "accuracy": 0.4788732394366197,
2994
+ "count": 213
2995
  }
2996
  }
2997
  },
2998
  "sub_M4": {
2999
+ "full_accuracy": 0.04,
3000
+ "digit_accuracy": 0.61,
3001
+ "n_examples": 100,
3002
  "per_subtask": {
3003
  "MD": {
3004
+ "accuracy": 0.985,
3005
+ "count": 200
3006
  },
3007
  "MB": {
3008
+ "accuracy": 0.96,
3009
+ "count": 100
3010
  },
3011
  "UB": {
3012
+ "accuracy": 0.47,
3013
+ "count": 100
3014
  },
3015
  "UD": {
3016
+ "accuracy": 0.29,
3017
+ "count": 300
3018
  }
3019
  }
3020
  },
3021
  "sub_M5": {
3022
+ "full_accuracy": 0.04,
3023
+ "digit_accuracy": 0.4942857142857143,
3024
+ "n_examples": 100,
3025
  "per_subtask": {
3026
  "MD": {
3027
  "accuracy": 1.0,
3028
+ "count": 100
3029
  },
3030
  "MB": {
3031
  "accuracy": 1.0,
3032
+ "count": 100
3033
  },
3034
  "UB": {
3035
+ "accuracy": 0.4,
3036
+ "count": 100
3037
  },
3038
  "UD": {
3039
+ "accuracy": 0.265,
3040
+ "count": 400
3041
  }
3042
  }
3043
  },
3044
  "sub_random": {
3045
+ "full_accuracy": 0.53,
3046
+ "digit_accuracy": 0.9192857142857143,
3047
  "n_examples": 200,
3048
  "per_subtask": {
3049
  "MD": {
3050
+ "accuracy": 0.96,
3051
+ "count": 600
3052
  },
3053
  "MB": {
3054
+ "accuracy": 0.947565543071161,
3055
+ "count": 267
3056
  },
3057
  "ME": {
3058
+ "accuracy": 0.9811320754716981,
3059
  "count": 53
3060
  },
3061
  "UB": {
3062
+ "accuracy": 0.8428246013667426,
3063
+ "count": 439
3064
  },
3065
  "UD": {
3066
+ "accuracy": 0.8780487804878049,
3067
+ "count": 41
3068
  }
3069
  }
3070
  },
3071
  "sub_B3": {
3072
+ "full_accuracy": 0.2,
3073
+ "digit_accuracy": 0.8371428571428572,
3074
+ "n_examples": 100,
3075
  "per_subtask": {
3076
  "MD": {
3077
+ "accuracy": 0.98,
3078
+ "count": 300
3079
  },
3080
  "MB": {
3081
+ "accuracy": 0.97,
3082
+ "count": 100
3083
  },
3084
  "UB": {
3085
+ "accuracy": 0.6598984771573604,
3086
+ "count": 197
3087
  },
3088
  "UD": {
3089
+ "accuracy": 0.6310679611650486,
3090
+ "count": 103
3091
  }
3092
  }
3093
  },
3094
  "sub_B4": {
3095
+ "full_accuracy": 0.02,
3096
+ "digit_accuracy": 0.7157142857142857,
3097
+ "n_examples": 100,
3098
  "per_subtask": {
3099
  "MD": {
3100
+ "accuracy": 0.99,
3101
+ "count": 200
3102
  },
3103
  "MB": {
3104
+ "accuracy": 0.97,
3105
+ "count": 100
3106
  },
3107
  "UB": {
3108
+ "accuracy": 0.5344129554655871,
3109
+ "count": 247
3110
  },
3111
  "UD": {
3112
+ "accuracy": 0.48366013071895425,
3113
+ "count": 153
3114
  }
3115
  }
3116
  },
3117
  "sub_B5": {
3118
+ "full_accuracy": 0.06,
3119
+ "digit_accuracy": 0.6771428571428572,
3120
+ "n_examples": 100,
3121
  "per_subtask": {
3122
  "MD": {
3123
  "accuracy": 1.0,
3124
+ "count": 100
3125
  },
3126
  "MB": {
3127
  "accuracy": 1.0,
3128
+ "count": 100
3129
  },
3130
  "UB": {
3131
+ "accuracy": 0.5906040268456376,
3132
+ "count": 298
3133
  },
3134
  "UD": {
3135
+ "accuracy": 0.48514851485148514,
3136
+ "count": 202
3137
  }
3138
  }
3139
  }
3140
  },
3141
  "summary": {
3142
+ "overall_accuracy": 0.3565384615384615,
3143
+ "digit_accuracy": 0.8012637362637363,
3144
+ "total_examples": 2600,
3145
  "n_splits": 24
3146
  }
3147
  }
add_sub_baseline_100K_1L2H256d/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0afe867ad343dff6b097dcd2a7666184ac3d7da5c44be1b7ae14dfcf95952c0d
3
  size 315072674
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:647cf0cb412885424f02571e9b561bfe0598c21caa2333555d8deeb5e3f232c8
3
  size 315072674
add_sub_baseline_100K_1L2H256d/train_config.json CHANGED
@@ -69,16 +69,20 @@
69
  "no_wandb": false,
70
  "n_params": 78691840,
71
  "run_name": "add_sub_baseline_100K_1L2H256d",
72
- "git_commit": "f447da529caceac8c7d256cbb2cd185cbc50feac",
73
- "timestamp": "2026-04-12T12:26:43.304885+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": "sft",
79
- "wandb_run_id": "qnb69zdv",
80
- "wandb_url": "https://wandb.ai/nlp_and_interpretability/sorl-arithmetic/runs/qnb69zdv",
81
- "final_accuracy": 0.33375,
82
- "sft_accuracy": 0.33375,
 
 
 
83
  "eval_method": "ArithmeticEvaluator"
84
  }
 
69
  "no_wandb": false,
70
  "n_params": 78691840,
71
  "run_name": "add_sub_baseline_100K_1L2H256d",
72
+ "git_commit": "f835493c19eb98267697007042c9d440cad2afbb",
73
+ "timestamp": "2026-04-15T12:55:07.216941+00:00",
74
  "tokenizer": "Qwen/Qwen3-0.6B",
75
  "dataset_repo": "thoughtworks/arithmetic-sorl-data",
76
  "dataset_config": "add_sub_6digit",
77
+ "train_dataset": "fixed_train/train_100K_seed42.pt",
78
  "model_repo": "thoughtworks/arithmetic-sorl",
79
  "trainer_version": "sft",
80
+ "wandb_run_id": "s4hx09ml",
81
+ "wandb_url": "https://wandb.ai/nlp_and_interpretability/sorl-arithmetic/runs/s4hx09ml",
82
+ "eval_final_dataset": "eval_sets/eval_add_sub_6d_N100_seed42.json",
83
+ "eval_epoch_dataset": "eval_sets/eval_add_sub_6d_N25_seed42.json",
84
+ "eval_hf_repo": "thoughtworks/arithmetic-sorl-data",
85
+ "final_accuracy": 0.3565384615384615,
86
+ "sft_accuracy": 0.3565384615384615,
87
  "eval_method": "ArithmeticEvaluator"
88
  }