基于稀疏MLP值向量的免训练真实性检测
Training-free Truthfulness Detection via Sparse MLP Value Vectors
AI总结：
提出TruthV方法，利用MLP值向量中稀疏子集的稳定方向性信号，无需训练即可检测LLM生成内容的真实性，在多个基准上超越现有免训练方法。
摘要：
大型语言模型（LLM）容易生成事实不正确的内容，这促使了从内部模型信号评估真实性的方法。虽然监督探测方法可能有效，但它们需要标记数据和分类器训练。最近的免训练方法避免了参数优化，但依赖于粗粒度的激活统计，这些统计对真实性相关信号如何在模型内产生提供的洞察有限。我们提出了一种在单个多层感知器（MLP）值向量级别操作的免训练方法。通过系统分析，我们发现尽管大多数值向量没有显示出有意义的信号，但一个稀疏子集表现出与内容真实性的稳定且方向一致的关联。利用这一观察，我们提出了\textbf{TruthV}，一种简单的推理方法，聚合这些值向量表达的偏好。TruthV只需要一个小的支持集来识别相关向量，并且不引入额外的模型参数或分类器权重。我们在从2B到13B的模型规模以及多个基准上评估了TruthV，包括问答、自然语言理解和幻觉评估。TruthV始终优于现有的免训练基线，表明LLM中与真实性相关的变化在MLP值向量级别以稀疏且结构化的方式被捕获。
英文摘要：
Large language models (LLMs) are prone to generating factually incorrect content, motivating methods for assessing truthfulness from internal model signals. While supervised probing approaches can be effective, they require labeled data and classifier training. Recent training-free methods avoid parameter optimization but rely on coarse activation statistics that provide limited insight into how truthfulness-related signals arise within the model. We present a training-free approach that operates at the level of individual multi-layer perceptron (MLP) value vectors. Through a systematic analysis, we find that although most value vectors show no meaningful signal, a sparse subset exhibits stable and directionally consistent correlations with content truthfulness. Leveraging this observation, we propose \textbf{TruthV}, a simple inference method that aggregates preferences expressed by these value vectors. TruthV requires only a small support set to identify relevant vectors and introduces no additional model parameters or classifier weights. We evaluate TruthV across model scales from 2B to 13B and multiple benchmarks, including question answering, natural language understanding, and hallucination evaluation. TruthV consistently outperforms existing training-free baselines, demonstrating that truthfulness-related variation in LLMs is captured in a sparse and structured manner at the level of MLP value vectors.
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。
