An Information-Theoretic Study of RLHF-Induced Uniformity in Large Language Model Outputs.

Nolan Chai, Tianqi Zhang, Alex Warstadt

Browse the full CoNLL paper archive.