Fetching the paper…

RM-Bench: Benchmarking Reward Models of Language Models with Subtlety and Style · Around