List your company

Reasoning on 2WikiMultiHopQA (OOD)

0Degeneration CountPrompt
-41.36237.82517796.18Feb 26, 2026

Evaluation results

MethodLinks
PromptModel Scale=1.7B2026.020544418907544
MBT-SModel Scale=4B2026.020460456491460
MBT-RModel Scale=4B2026.020649643692649
BaseModel Scale=0.6B2026.021569427719566
PromptModel Scale=0.6B2026.021555426681552
GRPOModel Scale=1.7B2026.0216374731,314634
BaseModel Scale=1.7B2026.022560397927555
MBT-SModel Scale=1.7B2026.024469450526459
BaseModel Scale=4B2026.0245013941,128491
MBT-RModel Scale=1.7B2026.028684655757664
PromptModel Scale=4B2026.02125354181,114504
MBT-RModel Scale=0.6B2026.0213682635764649
MBT-SModel Scale=0.6B2026.0214505454595469
RSModel Scale=1.7B2026.02336834081,432599
RSModel Scale=0.6B2026.0250703430957575
GRPOModel Scale=0.6B2026.02591,4099202,3591,261
GRPOModel Scale=4B2026.021218364262,274526
RSModel Scale=4B2026.021218153971,798505
gpt-oss-distillModel Scale=4B2026.021931,2716703,450780
TokenSkipModel Scale=4B2026.024261,6013903,361508
LIMOProModel Scale=0.6B2026.024581,5092982,157328
TokenSkipModel Scale=0.6B2026.026022,1824513,196644
LIMOProModel Scale=1.7B2026.026291,9533073,090331
gpt-oss-distillModel Scale=1.7B2026.026642,5577606,499873
gpt-oss-distillModel Scale=0.6B2026.027682,8497485,824903
LIMOProModel Scale=4B2026.028102,4343174,050346
TokenSkipModel Scale=1.7B2026.021,0343,2214257,257574