ds1000: by examples

Results Paper Code


Not solved by any model

There are 274 examples not solved by any model. Solving some of these can be a good signal that your model is indeed better than leading models if these are good problems.
104, 107, 108, 121, 122, 131, 142, 15, 157, 172, 173, 174, 181, 182, 183, 184, 195, 196, 197, 199, 200, 201, 202, 203, 204, 208, 209, 210, 211, 220, 221, 222, 225, 242, 244, 245, 269, 270, 272, 281, 282, 283, 284, 285, 286, 29, 328, 339, 362, 371, 372, 375, 379, 380, 385, 389, 39, 390, 394, 40, 408, 410, 42, 420, 421, 43, 44, 445, 45, 455, 46, 47, 475, 48, 486, 487, 488, 49, 509, 516, 520, 521, 585, 59, 6, 60, 612, 65, 666, 667, 668, 669, 67, 670, 671, 672, 673, 674, 675, 676, 677, 678, 679, 680, 681, 682, 683, 684, 685, 686, 687, 688, 689, 690, 691, 692, 693, 694, 695, 696, 697, 698, 699, 700, 701, 702, 703, 704, 705, 706, 707, 708, 709, 710, 726, 727, 73, 74, 745, 747, 75, 763, 764, 766, 775, 776, 779, 780, 789, 790, 794, 795, 796, 798, 808, 81, 810, 817, 818, 819, 820, 821, 822, 823, 824, 825, 826, 827, 831, 832, 833, 834, 835, 836, 837, 838, 839, 840, 841, 842, 843, 844, 845, 846, 847, 848, 849, 850, 851, 852, 853, 854, 855, 856, 857, 858, 859, 86, 860, 861, 862, 863, 864, 865, 866, 867, 868, 869, 870, 871, 872, 873, 874, 875, 876, 877, 878, 879, 88, 880, 881, 882, 886, 887, 888, 889, 890, 891, 892, 893, 894, 895, 896, 897, 898, 899, 9, 900, 901, 902, 903, 904, 905, 906, 907, 908, 909, 910, 911, 912, 913, 914, 915, 916, 917, 918, 919, 920, 921, 922, 923, 924, 925, 926, 927, 928, 929, 930, 931, 936, 937, 96, 987, 993

Problems solved by 1 model only

example_link model min_pass1_of_model
781 qwen2.5-coder-14b-instruct 0.432
66 qwen2.5-coder-14b-instruct 0.432
750 qwen2.5-coder-14b-instruct 0.432
338 qwen2.5-coder-14b-instruct 0.432
751 qwen2.5-coder-14b-instruct 0.432
58 qwen2.5-coder-14b-instruct 0.432
996 qwen3-14b 0.380
134 qwen3-14b 0.380
54 qwen3-14b 0.380
7 qwen3-14b 0.380
279 qwen3-14b 0.380
216 qwen3-14b 0.380
280 qwen3-14b 0.380
388 qwen3-14b 0.380
809 qwen3-14b 0.380
164 qwen3-14b 0.380
113 qwen2.5-coder-7b-instruct 0.357
112 google_gemma_3_12b_it 0.328
755 google_gemma_3_12b_it 0.328
446 google_gemma_3_12b_it 0.328
158 qwen3-8b 0.312
812 qwen3-8b 0.312
407 mistralai_ministral_8b_instruct_2410 0.247
161 qwen2.5-coder-3b-instruct 0.238
228 qwen2.5-coder-3b-instruct 0.238
227 qwen2.5-coder-3b-instruct 0.238
263 qwen2-7b-instruct 0.236
345 qwen1.5-14b-chat 0.205
635 google_gemma_3_4b_it 0.195
523 mistralai_mistral_7b_instruct_v0.3 0.182
800 deepseek_v2_lite_chat 0.174
468 deepseek_r1_distill_qwen_14b 0.161
953 deepseek_r1_distill_qwen_7b 0.148
596 qwen2.5-coder-1.5b-instruct 0.141
159 mistralai_mistral_7b_instruct_v0.2 0.136
447 google_gemma_7b_it 0.048

Suspect problems

These are 10 problems with the lowest correlation with the overall evaluation (i.e. better models tend to do worse on these. )

example_link pass1_of_ex tau
604 0.004 -0.171
802 0.011 -0.157
447 0.002 -0.119
603 0.038 -0.038
159 0.003 -0.024
649 0.303 -0.023
533 0.029 -0.016
596 0.002 -0.012
165 0.027 -0.010
953 0.002 0.000

Histogram of accuracies

Histogram of problems by the accuracy on each problem.

Histogram of difficulties

Histogram of problems by the minimum win rate to solve each problem.