List your company

Mathematical Reasoning on GSM8K (Acc, Peak, Dep)

87.92Accuracy (GSM8K)TokenSkip
73.04876.90980.7784.631Apr 4, 2026

Evaluation results

MethodLinks
TokenSkipModel Series=Qwen2.5-7B, Budget setting=true2026.0487.927751
LThinker++Model Series=Qwen2.5-7B, Budget setting=true2026.0487.874710.4
VanillaModel Series=Qwen2.5-7B, Budget setting=true2026.0487.627821.1
CoTModel Series=Qwen2.5-7B, Budget setting=true2026.0486.125130.1
CoTModel Series=Llama3.1-8B, Budget setting=true2026.0485.145500.2
LThinker*Model Series=Qwen2.5-7B, Budget setting=true2026.0484.614440.7
Distill-R1Model Series=Qwen2.5-7B, Budget setting=true2026.0481.888441.1
TokenSkipModel Series=Llama3.1-8B, Budget setting=true2026.0479.48381.2
VanillaModel Series=Llama3.1-8B, Budget setting=true2026.0479.387400.9
LThinker++Model Series=Llama3.1-8B, Budget setting=true2026.0477.695280.7
LThinker*Model Series=Llama3.1-8B, Budget setting=true2026.0476.725220.9
Distill-R1Model Series=Llama3.1-8B, Budget setting=true2026.0473.623950.1