自信的幻觉？通过邻域一致性诊断LLM的真实性
Illusions of Confidence? Diagnosing LLM Truthfulness via Neighborhood Consistency
AI总结：
本文提出邻域一致性信念(NCB)来评估LLM在上下文干扰下的信念鲁棒性，通过结构化测量提升模型在真实场景中的可靠性，并引入结构感知训练(SAT)减少知识脆性。
摘要：
随着大型语言模型(LLMs)在现实世界中的广泛应用，仅正确性已不够。可靠的部署需要在上下文扰动下维持真实信念。现有评估主要依赖点状置信度如自一致性，可能掩盖脆弱信念。我们证明即使完美自一致性回答在轻微上下文干扰下也会迅速崩溃。为解决这一差距，我们提出邻域一致性信念(NCB)，一种信念鲁棒性的结构度量，评估响应在概念邻域内的连贯性。为验证NCB的效率，我们引入新的认知压力测试协议，探测输出在上下文干扰下的稳定性。在多个LLM上的实验表明，高NCB数据相对更抗干扰。最后，我们提出结构感知训练(SAT)，通过优化上下文不变的信念结构，将长尾知识脆性减少约30%。代码可在https://github.com/zjunlp/belief获取。
英文摘要：
As Large Language Models (LLMs) are increasingly deployed in real-world settings, correctness alone is insufficient. Reliable deployment requires maintaining truthful beliefs under contextual perturbations. Existing evaluations largely rely on point-wise confidence like Self-Consistency, which can mask brittle belief. We show that even facts answered with perfect self-consistency can rapidly collapse under mild contextual interference. To address this gap, we propose Neighbor-Consistency Belief (NCB), a structural measure of belief robustness that evaluates response coherence across a conceptual neighborhood. To validate the efficiency of NCB, we introduce a new cognitive stress-testing protocol that probes outputs stability under contextual interference. Experiments across multiple LLMs show that the performance of high-NCB data is relatively more resistant to interference. Finally, we present Structure-Aware Training (SAT), which optimizes context-invariant belief structure and reduces long-tail knowledge brittleness by approximately 30%. Code is available at https://github.com/zjunlp/belief.
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。
