List your company

Scientific Reasoning on MMLU (Domain Specific Scores)

28.36Chemistry (Pro)RAISE
22.473624.001825.5327.0582Jun 10, 2025

Evaluation results

MethodLinks
RAISEStrategy Category=Ours2025.0628.3659.2751—
CoTStrategy Category=Direct2025.0625.4451.8849.5—
CoT+RAGStrategy Category=Direct+RAG2025.0625.1854.3943—
Least-to-MostStrategy Category=Decomposed2025.0624.5653.9745.4—
Step-Back+RAGStrategy Category=Decomposed+RAG2025.0623.0656.3443—
Least-to-Most+RAGStrategy Category=Decomposed+RAG2025.0622.9758.0246—
HyDEStrategy Category=Decomposed+RAG2025.0622.9757.8849—
Step-BackStrategy Category=Decomposed2025.0622.756.4943—
Base Model2025.09———48.16
EAPO2025.09———58.61
Self-Exploratory RL2025.09———55.47