B-Pref: Benchmark for Preference-based RL Please check out a new offifical codebase at https://github.com/rll-research/BPref