Rethinking the Role of Proxy Rewards in Language Model Alignment.

Sungdong Kim, Minjoon Seo

Browse the full EMNLP paper archive.