General Reasoning on Average of Reasoning Tasks
63.31Average AccuracyPASER
Evaluation Results
| Method | Links | |
|---|---|---|
| PASERQuantization=GPTQ 4 bits, Model=LLaMA2-7B2025.02 | 63.31 | |
| w/o trainingQuantization=w/o Quant, Model=LLaMA2-7B2025.02 | 62.91 | |
| PASERQuantization=RTN 4 bits, Model=LLaMA2-7B2025.02 | 62.58 | |
| NuggetsQuantization=GPTQ 4 bits, Model=LLaMA2-7B2025.02 | 61.96 | |
| IFDQuantization=GPTQ 4 bits, Model=LLaMA2-7B2025.02 | 61.42 | |
| NuggetsQuantization=RTN 4 bits, Model=LLaMA2-7B2025.02 | 61.23 | |
| Instruction MiningQuantization=GPTQ 4 bits, Model=LLaMA2-7B2025.02 | 60.68 | |
| IFDQuantization=RTN 4 bits, Model=LLaMA2-7B2025.02 | 60.49 | |
| Full DataQuantization=GPTQ 4 bits, Model=LLaMA2-7B2025.02 | 60.4 | |
| w/o trainingQuantization=GPTQ 4 bits, Model=LLaMA2-7B2025.02 | 60.21 | |
| Full DataQuantization=RTN 4 bits, Model=LLaMA2-7B2025.02 | 59.74 | |
| RandomQuantization=GPTQ 4 bits, Model=LLaMA2-7B2025.02 | 59.34 | |
| Instruction MiningQuantization=RTN 4 bits, Model=LLaMA2-7B2025.02 | 59.23 | |
| w/o trainingQuantization=RTN 4 bits, Model=LLaMA2-7B2025.02 | 59.05 | |
| RandomQuantization=RTN 4 bits, Model=LLaMA2-7B2025.02 | 57.48 | |
| + SPICEBackbone=Qwen3-8B-Base, #Data=20K2026.05 | 38.53 | |
| D2EvoBackbone=Qwen3-8B-Base, #Data=1.7K, Iteration=32026.05 | 36.13 | |
| + Full DataBackbone=Qwen3-8B-Base, #Data=19K2026.05 | 35.61 | |
| + AZRBackbone=Qwen3-8B-Base2026.05 | 35.6 | |
| + R-ZeroBackbone=Qwen3-8B-Base2026.05 | 35.28 | |
| Base ModelBackbone=Qwen3-8B-Base2026.05 | 33.54 | |
| + SPICEBackbone=Qwen3-4B-Base, #Data=20K2026.05 | 33.53 | |
| D2EvoBackbone=Qwen3-4B-Base, #Data=1.4K, Iteration=32026.05 | 32.16 | |
| + Full DataBackbone=Qwen3-4B-Base, #Data=19K2026.05 | 31.89 | |
| + R-ZeroBackbone=Qwen3-4B-Base2026.05 | 30.24 | |
| + AZRBackbone=Qwen3-4B-Base2026.05 | 29.33 | |
| D2EvoBackbone=Llama-3.1-8B, #Data=1.2K, Iteration=32026.05 | 29.07 | |
| + Full DataBackbone=Llama-3.1-8B, #Data=19K2026.05 | 28.39 | |
| Base ModelBackbone=Qwen3-4B-Base2026.05 | 27.96 | |
| Base ModelBackbone=Llama-3.1-8B2026.05 | 25.88 |