List your company

Multi-hop Question Answering on 2WikiMQA (test)

48.6Exact MatchQ-DREAM
22.629.3536.142.85May 31, 2025

Evaluation results

MethodLinks
Q-DREAMBackbone=ChatGPT2025.0548.662.1
IRCoTBackbone=ChatGPT2025.0541.955.2
SUREBackbone=ChatGPT2025.0532.838.1
Q-DREAMBackbone=Llama2-7B2025.0532.439.7
InstructRAGBackbone=Llama3-8B2025.0530.438.9
ChatQA2Backbone=Llama3-8B2025.052935.2
ChatGPTBackbone=GPT-3.5-turbo2025.0523.629.2