让AI自己辩论，判断上下文是否可信。
When to Trust Context: Self-Reflective Debates for Context Reliability
- 用自我反思辩论机制，区分模型知识与上下文冲突。
- 在冲突情境下准确率提升，对可靠上下文保持原有精度。
- 轻量设计，适合实际部署，适合需要可信推理的场景。
大语言模型常面临参数化知识与上下文输入之间的冲突，导致事实性不一致或幻觉。我们提出自反性上下文可靠性辩论（SR-DCR），一种轻量级框架，融合标记级自信心与非对称多智能体辩论以解决此类冲突。一名无上下文的评论者挑战基于给定文本辩护的防御者，由裁判模型评估辩论并判断上下文可靠性。最终答案通过结合裁决结果与模型置信度生成。在ClashEval基准上的实验表明，SR-DCR在面对误导性上下文时持续增强鲁棒性，同时在可信输入上保持准确率，优于经典辩论与仅依赖置信度的基线，且计算开销极低。代码已公开于https://github.com/smiles724/Self-Reflective-Debates。
原文摘要 · Abstract (English)
Large language models frequently encounter conflicts between their parametric knowledge and contextual input, often resulting in factual inconsistencies or hallucinations. We propose Self-Reflective Debate for Contextual Reliability (SR-DCR), a lightweight framework that integrates token-level self-confidence with an asymmetric multi-agent debate to adjudicate such conflicts. A critic, deprived of context, challenges a defender who argues from the given passage; a judge model evaluates the debate and determines the context's reliability. The final answer is selected by combining the verdict with model confidence. Experiments on the ClashEval benchmark demonstrate that SR-DCR consistently enhances robustness to misleading context while maintaining accuracy on trustworthy inputs, outperforming both classical debate and confidence-only baselines with minimal computational overhead. The code is available at https://github.com/smiles724/Self-Reflective-Debates.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。