Many of Your DPOs are Secretly One: Attempting Unification Through Mutual Information
원문 초록
Post-alignment of large language models (LLMs) is critical in improving their utility, safety, and alignment with human intentions. Direct preference optimisation (DPO) has become one of the most widely used algorithms for achieving this alignment, given its ability to optimise models based on human feedback directly. However, the vast number of DPO variants in the literature has made it increasingly difficult for researchers to navigate and fully grasp the connections between these approaches. This paper introduces a unifying framework inspired by mutual information, which proposes a new loss function with flexible priors. By carefully specifying these priors, we demonstrate that many existing algorithms, such as SimPO, TDPO, SparsePO, and others, can be derived from our framework. This unification offers a clearer and more structured approach, allowing researchers to understand the relationships between different DPO variants better. We aim to simplify the landscape of DPO algorithms, making it easier for the research community to gain insights and foster further advancements in LLM alignment. Ultimately, we hope our framework can be a foundation for developing more robust and interpretable alignment techniques.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
STRUCTSENSE: A Task-Agnostic Agentic Framework for Structured Information Extraction with Human-In-The-Loop Evaluation and Benchmarking
AI Research Agents for Machine Learning: Search, Exploration, and Generalization in MLE-bench
The Decrypto Benchmark for Multi-Agent Reasoning and Theory of Mind
Decide less, communicate more: On the construct validity of end-to-end fact-checking in medicine
From Emergence to Control: Probing and Modulating Self-Reflection in Language Models
A Navigation Framework Utilizing Vision-Language Models
토론
이 논문과 연결된 포럼 글이 아직 없습니다.