List your company

Question Answering on HotpotQA (F1, EM)

60.62F1 ScorePRAISE
23.835233.385142.93552.4849Apr 4, 2026

Evaluation results

MethodLinks
PRAISEApproach=Ours2026.0460.6247.17
R1-SearcherApproach=Agentic Search2026.0455.5543.31
Search-R1Approach=Agentic Search2026.0451.2439.77
StepSearchApproach=Process Reward2026.0450.0238.31
TIPSApproach=Process Reward2026.0445.9835.34
Search-o1Approach=Agentic Search2026.0444.8634.07
ReasonRAGApproach=Process Reward2026.0441.5331.24
Naive RAGApproach=Standard2026.0436.6927.45
Direct LLMApproach=Standard2026.0425.2517.6