search

wzhouad / Contra-OOD

Source code for paper "Contrastive Out-of-Distribution Detection for Pretrained Transformers", EMNLP 2021
MIT License
40 stars 9 forks source link

Problem of Table 2 #2

Closed wjczf123 closed 2 years ago

wjczf123 commented 2 years ago

The Table 2 seems confused. Why L_{margin} + MSP are not in green on SST2 dataset and in green on IMDb dataset? You write 'Results where the contrastive loss improves OOD detection on both evaluation metrics are highlighted in green.' And lower FAR95 is better.

wzhouad commented 2 years ago

The SST2 cell should be in green, and the IMDB cell shouldn't. Thank you for pointing this out.

wjczf123 commented 2 years ago

OK. Thank you for your reply.