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

Output Consistency on Subject and Property Perturbations

91Subject Agreement

GPT-4

2.625.5548.571.45Jan 22, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.01
9194
2025.01
8998
2025.01
7687
2025.01
7684
2025.01
7282
2025.01
6788
2025.01
5734
2025.01
5666
2025.01
5669
2025.01
5029
2025.01
4849
2025.01
4727
2025.01
4631
2025.01
4125
2025.01
3924
2025.01
3923
2025.01
3225
2025.01
3125
2025.01
3011
2025.01
2922
2025.01
2820
2025.01
109
2025.01
1019
2025.01
64