Mathematical Reasoning on MultiArith (Acc, Avg Tok, Avg Steps)
43.33AccuracyCoT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CoTModel=Llama-3.2-1B2026.03 | 43.33 | 28 | — | |
| CoTModel=Qwen2.5-1.5B2026.03 | 34.3 | 30.2 | — | |
| AdaAnchor (adaptive)Model=Llama-3.2-1B2026.03 | 32.44 | 2.57 | 3.5 | |
| iCoTModel=Llama-3.2-1B2026.03 | 30.84 | 2.12 | — | |
| AdaAnchor (adaptive)Model=Qwen2.5-1.5B2026.03 | 29.4 | 2.16 | 3.82 | |
| iCoTModel=Qwen2.5-1.5B2026.03 | 28.56 | 1.66 | — | |
| AdaAnchor (K=8)Model=Llama-3.2-1B2026.03 | 28.31 | 2.48 | 8 | |
| AdaAnchor (K=8)Model=Qwen2.5-1.5B2026.03 | 27.6 | 2.34 | 8 | |
| No CoTModel=Qwen2.5-1.5B2026.03 | 22.3 | 2.41 | — | |
| No CoTModel=Llama-3.2-1B2026.03 | 20.56 | 2.08 | — |