Deterministic Inference Evaluation on AIME'24
0Avg Max Probability DivergenceBIO+TBIK
Evaluation Results
| Method | Links | |
|---|---|---|
| BIO+TBIKModel=Qwen3-8B2025.11 | 0 | |
| BIO+TBIKModel=Mistral-7B-Instruct2025.11 | 0 | |
| BIO+TBIKModel=Llama-3.1-8B-Instruct2025.11 | 0 | |
| BIO+TBIKModel=Qwen3-32B2025.11 | 0 | |
| BIOModel=Qwen3-32B2025.11 | 7.79 | |
| BF16Model=Qwen3-32B2025.11 | 8.01 | |
| BIOModel=Qwen3-8B2025.11 | 9.1 | |
| BF16Model=Qwen3-8B2025.11 | 10.01 | |
| BIOModel=Mistral-7B-Instruct2025.11 | 14.1 | |
| BF16Model=Mistral-7B-Instruct2025.11 | 17.37 | |
| BF16Model=Llama-3.1-8B-Instruct2025.11 | 26.48 | |
| BIOModel=Llama-3.1-8B-Instruct2025.11 | 27.54 |