密度脊选择性预测：校准标签稀缺下的大语言模型与视觉语言模型幻觉检测
Density Ridge Selective Prediction for LLM and VLM Hallucination Detection under Calibration Label Scarcity
AI总结：
针对校准标签稀缺时大语言模型和视觉语言模型的幻觉检测问题，提出基于核密度估计的密度脊方法，利用隐藏状态生成轨迹的六维运动特征图构建响应流形，通过到最近脊顶点的欧氏距离评分，在标签稀缺协议下AUROC提升5-20点。
摘要：
大语言模型和视觉语言模型中的幻觉检测日益被框架化为选择性预测，其中检测器分配置信度分数并在置信度低时弃权。无监督采样检测器（Semantic Entropy, EigenScore）避免标签但质量停滞，而有监督探针（SAPLMA）获得更强的分布内分数，但在校准标签稀缺时性能急剧下降。我们将大语言模型的响应流形恢复为基于隐藏状态生成轨迹的六维运动特征图的核密度估计的密度脊。测试生成通过其投影特征点到最近脊顶点的欧氏距离的负值进行评分，从而得到随机输出分布的低维几何骨架。我们在七个问答基准（HaluEval-QA, TriviaQA, GSM8K, POPE, ScienceQA, A-OKVQA）上，使用九个文本和视觉大语言模型，在刻意标签稀缺协议（$n_{\ ext{cal}}{=}200$ 查询，$N{=}5$ 生成）下，与Semantic Entropy、SAR、EigenScore、SAPLMA和对数概率进行评估。我们的基于脊的分数在AUROC上以5-20个百分点的优势获胜，同时在校准标签稀缺下表现出温和的性能下降。
英文摘要：
Hallucination detection in large language and vision-language models is increasingly framed as selective prediction, where a detector assigns a confidence score and abstains when confidence is low. Unsupervised sampling detectors (Semantic Entropy) avoid labels but plateau in quality, while supervised probes attain stronger in-distribution scores yet degrade sharply when calibration labels are scarce. We recover the response manifold of an LLM as the density ridge of a kernel density estimate built on a six-dimensional kinematic feature map of hidden state generation trajectories. A test generation is scored by the negated Euclidean distance from its projected feature point to the nearest ridge vertex, yielding a low-dimensional geometric skeleton of the stochastic output distribution. We evaluate against Semantic Entropy, topological methods, and log-probability on six QA benchmarks (HaluEval-QA, TriviaQA, GSM8K, POPE, ScienceQA, A-OKVQA) using eight text and vision LLMs in a deliberately label-scarce protocol ($n_{\text{cal}}{=}200$ queries, $N{=}5$ generations). Our ridge-based score beats on AUROC with 5-20 points gain, while demonstrating tempered degradation under calibration-label scarcity.
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。
