Graduate-level Reasoning on GPQA-D (Score)
61.1ScoreMTI
Evaluation Results
| Method | Links | |
|---|---|---|
| MTIModel=Qwen3-8B, τ=1.52025.10 | 61.1 | |
| MTIModel=Qwen3-8B, τ=0.52025.10 | 60.6 | |
| MTIModel=Qwen3-8B, τ=1.02025.10 | 60.1 | |
| CGRSModel=Qwen3-8B, τ=-2025.10 | 59.8 | |
| DEERModel=Qwen3-8B, τ=-2025.10 | 59.3 | |
| TALEModel=Qwen3-8B, τ=-2025.10 | 59.1 | |
| DynasorModel=Qwen3-8B, τ=-2025.10 | 57.7 | |
| NoThinkingModel=Qwen3-8B, τ=-2025.10 | 54.2 | |
| MTIModel=DeepSeek-R1-7B, τ=0.52025.10 | 51.52 | |
| MTIModel=DeepSeek-R1-7B, τ=1.02025.10 | 48.5 | |
| MTIModel=DeepSeek-R1-7B, τ=1.52025.10 | 40.9 | |
| NoThinkingModel=DeepSeek-R1-7B, τ=-2025.10 | 37.9 | |
| TALEModel=DeepSeek-R1-7B, τ=-2025.10 | 36.2 | |
| DEERModel=DeepSeek-R1-7B, τ=-2025.10 | 33.1 | |
| CGRSModel=DeepSeek-R1-7B, τ=-2025.10 | 32.8 | |
| DynasorModel=DeepSeek-R1-7B, τ=-2025.10 | 22.2 |