List your company

Question Answering on PopQA (F1, SBERT, Acc)

50.8F1 ScoreCRITIC-R1
9.92820.53931.1541.761May 28, 2026

Evaluation results

MethodLinks
CRITIC-R1Method Category=Reflective / Critique Methods2026.0550.867.369.2
Search-R1Method Category=Reasoning and Retrieval Baselines, Status=criticized base RAG model2026.0550.166.866.1
Self-RefineMethod Category=Reflective / Critique Methods2026.0546.334.548.3
Align-RAGMethod Category=Reflective / Critique Methods2026.0541.93647.3
Self-RAGMethod Category=Reflective / Critique Methods2026.052141.967.3
CoTMethod Category=Reasoning and Retrieval Baselines2026.0516.847.933.8
VanillaMethod Category=Base LLM Generation2026.0511.732.327.9
Naive-RAGMethod Category=Reasoning and Retrieval Baselines2026.0511.533.244.6