Formality is Favored: Unraveling the Learning Preferences of Large Language Models on Data with Conflicting Knowledge.

Jiahuan Li, Yiqing Cao, Shujian Huang, Jiajun Chen

Browse the full EMNLP paper archive.