Mathematical Reasoning on Math-500 (Acc., Avg. Δ)
95.2AccuracyPaTaRM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PaTaRMPolicy Model=Qwen3-8B2025.10 | 95.2 | — | |
| Rule-based RewardPolicy Model=Qwen3-8B2025.10 | 95 | — | |
| Skywork-Llama-3.1-8BPolicy Model=Qwen3-8B2025.10 | 93.6 | — | |
| Base ModelPolicy Model=Qwen3-8B2025.10 | 90 | — | |
| PaTaRMPolicy Model=Qwen3-0.6B2025.10 | 78 | — | |
| Rule-based RewardPolicy Model=Qwen3-0.6B2025.10 | 76.4 | — | |
| Skywork-Llama-3.1-8BPolicy Model=Qwen3-0.6B2025.10 | 74.2 | — | |
| Base ModelPolicy Model=Qwen3-0.6B2025.10 | 72.2 | — | |
| DECOVECModel=Qwen2-7B, Protocol=Few-Shot (BM25)2026.04 | 48.83 | 0.59 | |
| DECOVECModel=Qwen2-7B, Protocol=Few-Shot (KATE)2026.04 | 43.36 | 5.49 | |
| DECOVECModel=Qwen2-7B, Protocol=Few-Shot (Rand. Std.)2026.04 | 41.02 | 1.38 | |
| DECOVECModel=Gemma-2-9B, Protocol=Few-Shot (BM25)2026.04 | 37.89 | 2.35 | |
| DECOVECModel=Llama-3-8B, Protocol=Few-Shot (KATE)2026.04 | 37.11 | 2.96 | |
| DECOVECModel=Qwen2-7B, Protocol=Few-Shot (Rand. Ext.)2026.04 | 36.33 | 2.15 | |
| DECOVECModel=Gemma-2-9B, Protocol=Few-Shot (KATE)2026.04 | 35.94 | 2.36 | |
| DECOVECModel=Llama-3-8B, Protocol=Few-Shot (BM25)2026.04 | 33.2 | 1.18 | |
| DECOVECModel=Gemma-2-9B, Protocol=Few-Shot (Rand. Ext.)2026.04 | 26.56 | 1.38 | |
| DECOVECModel=Llama-3-8B, Protocol=Few-Shot (Rand. Ext.)2026.04 | 21.88 | 2.16 | |
| Zero-ShotModel=Qwen2-7B, Protocol=Zero-Shot2026.04 | 21.09 | — | |
| DECOVECModel=Gemma-2-9B, Protocol=Few-Shot (Rand. Std.)2026.04 | 21.09 | 5.49 | |
| DECOVECModel=Qwen2-1.5B, Protocol=Few-Shot (BM25)2026.04 | 20.7 | 4.9 | |
| DECOVECModel=Qwen2-1.5B, Protocol=Few-Shot (KATE)2026.04 | 19.92 | 5.5 | |
| DECOVECModel=Llama-3-8B, Protocol=Few-Shot (Rand. Std.)2026.04 | 18.75 | 2.55 | |
| DECOVECModel=Qwen2-1.5B, Protocol=Few-Shot (Rand. Ext.)2026.04 | 17.97 | 0.99 | |
| DECOVECModel=Qwen2-1.5B, Protocol=Few-Shot (Rand. Std.)2026.04 | 17.58 | 2.56 | |
| DECOVECModel=Yi-6B, Protocol=Few-Shot (BM25)2026.04 | 17.58 | 1.96 | |
| DECOVECModel=Yi-6B, Protocol=Few-Shot (KATE)2026.04 | 17.19 | 3.54 | |
| DECOVECModel=Llama-2-7B, Protocol=Few-Shot (BM25)2026.04 | 12.5 | 2.74 | |
| DECOVECModel=Qwen2-0.5B, Protocol=Few-Shot (BM25)2026.04 | 10.55 | 0.79 | |
| DECOVECModel=Llama-2-7B, Protocol=Few-Shot (KATE)2026.04 | 9.77 | 3.91 | |
| DECOVECModel=Qwen2-0.5B, Protocol=Few-Shot (Rand. Std.)2026.04 | 8.59 | 2.37 | |
| DECOVECModel=Qwen2-0.5B, Protocol=Few-Shot (KATE)2026.04 | 8.59 | 4.72 | |
| DECOVECModel=Qwen2-0.5B, Protocol=Few-Shot (Rand. Ext.)2026.04 | 7.03 | 2.17 | |
| DECOVECModel=Yi-6B, Protocol=Few-Shot (Rand. Std.)2026.04 | 5.47 | 0.39 | |
| Zero-ShotModel=Gemma-2-9B, Protocol=Zero-Shot2026.04 | 5.47 | — | |
| DECOVECModel=Yi-6B, Protocol=Few-Shot (Rand. Ext.)2026.04 | 5.08 | 2.45 | |
| Zero-ShotModel=Yi-6B, Protocol=Zero-Shot2026.04 | 3.52 | — | |
| DECOVECModel=Llama-2-7B, Protocol=Few-Shot (Rand. Ext.)2026.04 | 2.73 | 1.57 | |
| DECOVECModel=Llama-2-7B, Protocol=Few-Shot (Rand. Std.)2026.04 | 1.95 | 1.57 | |
| Zero-ShotModel=Qwen2-0.5B, Protocol=Zero-Shot2026.04 | 1.17 | — | |
| Zero-ShotModel=Qwen2-1.5B, Protocol=Zero-Shot2026.04 | 0.78 | — | |
| Zero-ShotModel=Llama-2-7B, Protocol=Zero-Shot2026.04 | 0 | — | |
| Zero-ShotModel=Llama-3-8B, Protocol=Zero-Shot2026.04 | 0 | — |