List your company

Question Answering on HotpotQA subset of 500 questions (dev)

31.4EM (ReAct)FireAct
14.13618.61823.127.582Oct 9, 2023

Evaluation results

MethodLinks
FireActBackbone=GPT-3.52023.1031.439.27.825
FireActBackbone=CodeLlama-34B2023.1022.227.85.625
FireActBackbone=Llama-2-13B2023.1021.234.413.162
FireActBackbone=CodeLlama-13B2023.1020.8298.239
FireActBackbone=CodeLlama-7B2023.1017.427.810.460
FireActBackbone=Llama-2-7B2023.1014.826.211.477