Linear Alignment: A Closed-form Solution for Aligning Human Preferences without Tuning and Feedback
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911939496509440 |
|---|---|
| author | Gao, Songyang Ge, Qiming Shen, Wei Dou, Shihan Ye, Junjie Wang, Xiao Zheng, Rui Zou, Yicheng Chen, Zhi Yan, Hang Zhang, Qi Lin, Dahua |
| author_facet | Gao, Songyang Ge, Qiming Shen, Wei Dou, Shihan Ye, Junjie Wang, Xiao Zheng, Rui Zou, Yicheng Chen, Zhi Yan, Hang Zhang, Qi Lin, Dahua |
| contents | The success of AI assistants based on Language Models (LLMs) hinges on Reinforcement Learning from Human Feedback (RLHF) to comprehend and align with user intentions. However, traditional alignment algorithms, such as PPO, are hampered by complex annotation and training requirements. This reliance limits the applicability of RLHF and hinders the development of professional assistants tailored to diverse human preferences. In this work, we introduce \textit{Linear Alignment}, a novel algorithm that aligns language models with human preferences in one single inference step, eliminating the reliance on data annotation and model training. Linear alignment incorporates a new parameterization for policy optimization under divergence constraints, which enables the extraction of optimal policy in a closed-form manner and facilitates the direct estimation of the aligned response. Extensive experiments on both general and personalized preference datasets demonstrate that linear alignment significantly enhances the performance and efficiency of LLM alignment across diverse scenarios. Our code and dataset is published on \url{https://github.com/Wizardcoast/Linear_Alignment.git}. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_11458 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Linear Alignment: A Closed-form Solution for Aligning Human Preferences without Tuning and Feedback Gao, Songyang Ge, Qiming Shen, Wei Dou, Shihan Ye, Junjie Wang, Xiao Zheng, Rui Zou, Yicheng Chen, Zhi Yan, Hang Zhang, Qi Lin, Dahua Computation and Language The success of AI assistants based on Language Models (LLMs) hinges on Reinforcement Learning from Human Feedback (RLHF) to comprehend and align with user intentions. However, traditional alignment algorithms, such as PPO, are hampered by complex annotation and training requirements. This reliance limits the applicability of RLHF and hinders the development of professional assistants tailored to diverse human preferences. In this work, we introduce \textit{Linear Alignment}, a novel algorithm that aligns language models with human preferences in one single inference step, eliminating the reliance on data annotation and model training. Linear alignment incorporates a new parameterization for policy optimization under divergence constraints, which enables the extraction of optimal policy in a closed-form manner and facilitates the direct estimation of the aligned response. Extensive experiments on both general and personalized preference datasets demonstrate that linear alignment significantly enhances the performance and efficiency of LLM alignment across diverse scenarios. Our code and dataset is published on \url{https://github.com/Wizardcoast/Linear_Alignment.git}. |
| title | Linear Alignment: A Closed-form Solution for Aligning Human Preferences without Tuning and Feedback |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2401.11458 |