Reasoning on MuSiQue (OOD)
0Degeneracy ScoreBase
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BaseModel Scale=0.6B2026.02 | 0 | 731 | 575 | 785 | 731 | |
| PromptModel Scale=0.6B2026.02 | 0 | 697 | 551 | 744 | 697 | |
| MBT-SModel Scale=4B2026.02 | 0 | 471 | 449 | 514 | 471 | |
| BaseModel Scale=1.7B2026.02 | 1 | 1,186 | 849 | 1,447 | 1,173 | |
| GRPOModel Scale=1.7B2026.02 | 1 | 1,379 | 1,024 | 1,801 | 1,366 | |
| MBT-RModel Scale=1.7B2026.02 | 1 | 728 | 686 | 771 | 715 | |
| MBT-RModel Scale=4B2026.02 | 1 | 724 | 688 | 784 | 711 | |
| PromptModel Scale=1.7B2026.02 | 2 | 1,134 | 801 | 1,386 | 1,108 | |
| MBT-SModel Scale=1.7B2026.02 | 2 | 521 | 462 | 582 | 494 | |
| BaseModel Scale=4B2026.02 | 2 | 1,368 | 1,015 | 1,933 | 1,342 | |
| MBT-RModel Scale=0.6B2026.02 | 5 | 750 | 658 | 804 | 684 | |
| PromptModel Scale=4B2026.02 | 5 | 1,324 | 990 | 1,789 | 1,259 | |
| MBT-SModel Scale=0.6B2026.02 | 6 | 561 | 451 | 633 | 481 | |
| RSModel Scale=0.6B2026.02 | 12 | 943 | 578 | 1,049 | 784 | |
| GRPOModel Scale=0.6B2026.02 | 16 | 2,167 | 1,488 | 2,544 | 1,963 | |
| RSModel Scale=1.7B2026.02 | 19 | 1,533 | 934 | 1,981 | 1,286 | |
| GRPOModel Scale=4B2026.02 | 73 | 2,433 | 1,126 | 4,004 | 1,488 | |
| RSModel Scale=4B2026.02 | 133 | 3,097 | 1,048 | 4,675 | 1,369 | |
| TokenSkipModel Scale=0.6B2026.02 | 174 | 3,220 | 688 | 3,619 | 928 | |
| gpt-oss-distillModel Scale=4B2026.02 | 285 | 5,437 | 1,443 | 10,489 | 1,784 | |
| LIMOProModel Scale=0.6B2026.02 | 386 | 5,697 | 479 | 7,430 | 552 | |
| TokenSkipModel Scale=1.7B2026.02 | 471 | 7,437 | 942 | 9,889 | 1,306 | |
| gpt-oss-distillModel Scale=0.6B2026.02 | 584 | 9,465 | 1,555 | 14,329 | 2,040 | |
| gpt-oss-distillModel Scale=1.7B2026.02 | 585 | 9,322 | 1,454 | 16,290 | 1,835 | |
| LIMOProModel Scale=1.7B2026.02 | 585 | 8,393 | 516 | 11,539 | 610 | |
| LIMOProModel Scale=4B2026.02 | 590 | 8,478 | 560 | 12,193 | 634 | |
| TokenSkipModel Scale=4B2026.02 | 651 | 9,932 | 1,040 | 13,556 | 1,514 |