conftrace_

offline preference dataset

1 paper

Explore in graph

Co-occurring keywords · papers mentioning both

behavior llm (1) direct alignment from preference (1) online alignment (1) large language model (1) behavior model (1) trust region (1) language model alignment (1) preference optimization (1) preference alignment (1) online learning (1)

Papers

BPO: Staying Close to the Behavior LLM Creates Better Online LLM Alignment EMNLP 2024