List your company

Science Question Answering on GPQA Main

0.361AccuracyHD + Confidence
0.26220.287850.31350.33915Jan 9, 2026

Evaluation results

MethodLinks
HD + ConfidenceModel=Qwen-2.5-7B-Instruct2026.010.361
HD + Learn2AggModel=Qwen-2.5-7B-Instruct2026.010.353
ConfidenceModel=Llama-3.1-8B-Instruct2026.010.352
HD + ConfidenceModel=Llama-3.1-8B-Instruct2026.010.352
ConfidenceModel=Qwen-2.5-7B-Instruct2026.010.35
HD + Learn2AggModel=Llama-3.1-8B-Instruct2026.010.35
Debate 5 × 5Model=Qwen-2.5-7B-Instruct2026.010.339
Majority VoteModel=Qwen-2.5-7B-Instruct2026.010.335
High DiversityModel=Qwen-2.5-7B-Instruct2026.010.334
High DiversityModel=Llama-3.1-8B-Instruct2026.010.321
Debate 5 × 5Model=Llama-3.1-8B-Instruct2026.010.316
Single ModelModel=Qwen-2.5-7B-Instruct2026.010.298
Majority VoteModel=Llama-3.1-8B-Instruct2026.010.29
Single ModelModel=Llama-3.1-8B-Instruct2026.010.266