Karma

LLMs explore preference alignment and failure mitigation techniques

Researchers are exploring new methods for aligning large language models (LLMs) with human preferences and mitigating specific failure modes. One approach uses Direct Preference Optimization (DPO) to reduce text degener…