List your company

Question Answering on AmbigQA (test)

65.8Correctness (%)ConflictRAG
44.37649.93855.561.062May 17, 2026

Evaluation results

MethodLinks
ConflictRAGSeeds=32026.0565.846.368.2
CoT DetectionTemperature=02026.0560.543.241.6
NLI-FilterTemperature=02026.0555.140.534.5
CRAGTemperature=02026.0551.238.123.6
RAG+RerankTemperature=02026.0549.136.221.9
Standard RAGTemperature=02026.0546.834.520.8
Self-RAGTemperature=02026.0545.233.121.2