Adaptive Token Biaser：通过偏置关键实体进行知识编辑
Adaptive Token Biaser: Knowledge Editing via Biasing Key Entities
AI总结：
针对大语言模型参数知识易过时的问题，提出新型解码技术ATBias，通过偏置知识相关token的logits增强上下文编辑效果，性能提升最高达32.3%且延迟减半，可低成本广泛应用。
摘要：
大语言模型（LLM）记忆的参数化知识会快速过时。上下文编辑（ICE）是目前更新LLM知识最有效的方法。近期的研究进展通过修改解码策略来增强ICE，无需改变模型内部结构或调整外部提示。然而，这种增强作用于整个序列生成过程，涵盖大量非关键token。本研究提出了自适应Token偏置器（ATBias），这是一种旨在增强ICE的新型解码技术。它在解码过程中聚焦于与知识关联度最高的token，通过匹配与新知识、参数化知识相关的关键实体来偏置这些token的logits。实验结果表明，ATBias显著提升了ICE性能，相比当前最优的ICE方法最高提升32.3%，同时仅产生一半的延迟。ATBias不仅提升了ICE的知识编辑能力，还能以极低的成本广泛应用于各类LLM。
英文摘要：
The parametric knowledge memorized by large language models (LLMs) becomes outdated quickly. In-context editing (ICE) is currently the most effective method for updating the knowledge of LLMs. Recent advancements involve enhancing ICE by modifying the decoding strategy, obviating the need for altering internal model structures or adjusting external prompts. However, this enhancement operates across the entire sequence generation, encompassing a plethora of non-critical tokens. In this work, we introduce $\textbf{A}$daptive $\textbf{T}$oken $\textbf{Bias}$er ($\textbf{ATBias}$), a new decoding technique designed to enhance ICE. It focuses on the tokens that are mostly related to knowledge during decoding, biasing their logits by matching key entities related to new and parametric knowledge. Experimental results show that ATBias significantly enhances ICE performance, achieving up to a 32.3% improvement over state-of-the-art ICE methods while incurring only half the latency. ATBias not only improves the knowledge editing capabilities of ICE but can also be widely applied to LLMs with negligible cost.
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。
