基于神经老虎机的在线个性化白盒LLM生成
Online Personalizing White-box LLMs Generation with Neural Bandits
AI总结：
研究提出基于神经老虎机的在线方法，通过用户反馈动态优化软指令嵌入，提升白盒LLM个性化文本生成效果，在新闻标题生成任务中ROUGE分数提升达62.9%。
摘要：
LLM个性化内容生成的出现带来了新挑战：如何高效调整文本以满足个人偏好，而无需为每个用户创建独特模型这种不可持续的需求。本研究引入一种创新的在线方法，采用神经老虎机算法基于用户反馈动态优化软指令嵌入，增强白盒LLM开放式文本生成的个性化。通过对各种任务的严格实验，我们证明相比基线策略有显著性能提升。特别是NeuralTS在个性化新闻标题生成中取得大幅提升，最佳ROUGE分数提升达62.9%，LLM-agent评估相比基线提升达2.76%。
英文摘要：
The advent of personalized content generation by LLMs presents a novel challenge: how to efficiently adapt text to meet individual preferences without the unsustainable demand of creating a unique model for each user. This study introduces an innovative online method that employs neural bandit algorithms to dynamically optimize soft instruction embeddings based on user feedback, enhancing the personalization of open-ended text generation by white-box LLMs. Through rigorous experimentation on various tasks, we demonstrate significant performance improvements over baseline strategies. NeuralTS, in particular, leads to substantial enhancements in personalized news headline generation, achieving up to a 62.9% improvement in terms of best ROUGE scores and up to 2.76% increase in LLM-agent evaluation against the baseline.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
