发表机构
Crestwood Preparatory College; St. Theresa of Lisieux Catholic High School; University of Toronto（克雷斯特伍德预备学院; 圣特蕾莎·利雪天主高中; 多伦多大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对临床问答中LLM幻觉问题，提出ALTAS，利用单次前向传播的终端熵和晚期层线性度，按问题在贪心解码与轨迹校正间路由，在保持临床基准不伤害的同时提升真实性。
AI 中文摘要
大型语言模型（LLMs）在临床问答中常因其幻觉倾向而被认为不安全。检索增强、微调和外部验证器需要临床治理机构批准的新基础设施，并可能增加延迟或额外的模型调用。推理时校正利用模型内部的逻辑信号，但固定的变换未必适用于每个问题。一个在真实性压力测试上提高约十个百分点准确率的校正器，在临床多项选择基准上却收益甚微，因为指令微调将输出概率集中在单一答案上，导致终端熵较低。我们提出ALTAS，它通过一次前向传播读取终端熵和晚期层线性度（$R^2$），以在贪心解码和晚期层轨迹校正之间为每个问题做出选择。无需训练分类器、探针或头；路由器基于候选答案的逻辑值运行，并增加6.5%的延迟开销。应用于每个问题时，该校正使TruthfulQA在3B和8B规模上分别比贪心解码提高11.4和10.0个百分点（$p<10^{-10}$）。按问题门控后，ALTAS保留8.3至9.5个百分点的增益，同时使MedQA、PubMedQA和MedHallu保持在“不伤害”的一个百分点范围内，与贪心解码相比无统计学显著差异。该方法通过了冻结阈值、评分规则和领域标签的验证扫描。
英文摘要
Large language models (LLMs) are often deemed unsafe for clinical question answering because of their tendency to hallucinate. Retrieval augmentation, fine-tuning, and external verifiers require new infrastructure that clinical governance must approve and may add latency or extra model calls. Inference-time correction uses the model's internal logit signals, but a fixed transformation need not suit every question. A corrector that improves accuracy by about ten percentage points on a truthfulness stress test yields negligible gains on clinical multiple-choice benchmarks, where instruction tuning concentrates output probability on one answer and leaves low terminal entropy. We introduce ALTAS, which reads terminal entropy and late-layer linearity ($R^2$) from one forward pass to choose per question between greedy decoding and late-layer trajectory correction. No classifier, probe, or head is trained; the router operates on candidate-answer logits and adds 6.5% latency overhead. Applied to every question, the correction improves TruthfulQA over greedy at 3B and 8B by 11.4 and 10.0 percentage points, respectively ($p<10^{-10}$). Gated per question, ALTAS retains gains of 8.3 to 9.5 percentage points while keeping MedQA, PubMedQA, and MedHallu within a one-percentage-point do-no-harm band, with no statistically significant differences from greedy. The method passes verification sweeps over frozen thresholds, the scoring rule, and the domain label.
