List your company

Question Answering on HotpotQA (Accuracy and Latency)

32AccuracyToT
20.5623.5326.529.47Jun 13, 2024

Evaluation results

MethodLinks
ToTBackbone=LLAMA2-13B2024.06321,271
TS-SFTBackbone=LLAMA2-13B2024.0630.365.5
CPOBackbone=LLAMA2-13B2024.0630.363.8
ToTBackbone=Mistral-7B2024.06304,698
CPOBackbone=Mistral-7B2024.0629.456.9
CoTBackbone=LLAMA2-13B2024.062965.2
TS-SFTBackbone=Mistral-7B2024.0628.656.2
CoTBackbone=Mistral-7B2024.062858.4
CPOBackbone=LLAMA2-7B2024.062441.1
ToTBackbone=LLAMA2-7B2024.06231,100.7
TS-SFTBackbone=LLAMA2-7B2024.0622.744.8
CoTBackbone=LLAMA2-7B2024.062145.5