List your company

Question Answering on HotpotQA (F1, SBERT, Acc)

47.1F1 ScoreAlign-RAG
6.33216.91627.538.084May 28, 2026

Evaluation results

MethodLinks
Align-RAGMethod Category=Reflective / Critique Methods2026.0547.131.343.6
CRITIC-R1Method Category=Reflective / Critique Methods2026.0546.365.450.6
Search-R1Method Category=Reasoning and Retrieval Baselines, Status=criticized base RAG model2026.0544.564.448.6
Self-RefineMethod Category=Reflective / Critique Methods2026.0535.645.736.6
Self-RAGMethod Category=Reflective / Critique Methods2026.0519.642.138.4
CoTMethod Category=Reasoning and Retrieval Baselines2026.0513.630.633.6
Naive-RAGMethod Category=Reasoning and Retrieval Baselines2026.058.332.132.5
VanillaMethod Category=Base LLM Generation2026.057.929.433.9