Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Instruction Following on Vicuna benchmark

8.09GPT-4 Evaluation Score

llama2 → CP → FT + chat vector

4.47085.41046.357.2896Oct 7, 2023
Updated 3mo ago

Evaluation Results

MethodLinks
2023.10
8.09
2023.10
8.02
2023.10
7.89
2023.10
7.86
2023.10
7.58
2023.10
7.47
2023.10
7.37
2023.10
7.07
2023.10
7.06
2023.10
7.03
2023.10
6.7
2023.10
6.46
2023.10
6.13
2023.10
6.04
2023.10
5.89
2023.10
5.5
2023.10
5.06
2023.10
4.61