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

Instruction Following on Vicuna benchmark zero-shot

119.4Pairwise Score (ChatGPT vs Sys)

GPT-4

22.57647.71372.8597.987May 23, 2023
Updated 1mo ago

Evaluation Results

MethodLinks
2023.05
119.4110.1114.52.6
2023.05
96.7101.999.34.4
2023.05
96.599.297.84.4
2023.05
93.296.494.84.1
2023.05
91.298.794.94.5
2023.05
91.298.794.94.5
2023.05
87.393.490.45.2
2023.05
84.189.8875.4
2023.05
67.279.773.64.2
2023.05
63.876.769.44.2
2023.05
6377.970.74.3
2023.05
57.371.264.45
2023.05
55.569.162.54.7
2023.05
50.669.860.54.2
2023.05
49.269.359.54.7
2023.05
44.96253.65.2
2023.05
3860.549.14.6
2023.05
3759.648.44.6
2023.05
33.356.144.84
2023.05
32.461.2473.6
2023.05
26.349.7383.9