arXiv cs.CL· Fardin Ahsan Sakib, Ziwei Zhu, Ozlem Uzuner·· 12 小时前AI 评分34
Trajectory-Guided Steering：缓解临床分类中的虚假特征偏差
Trajectory-Guided Steering: Mitigating Spurious Feature Bias in Clinical Classification
AI 导读
研究者提出 Trajectory-Guided Steering（TGS），仅依据模型预测轨迹判断是否纠正及纠正方向，无需反事实数据或虚假特征先验。在 i2b2 的 16 个临床分类任务上，TGS 修复 17.7% 的错误，仅破坏 1.3% 的正确预测，所有测试超参数设置下均为净正收益。
正文 · 原文
Abstract:Large language models rely on spurious features in clinical classification: swapping a single gender token in clinical notes alters gender-irrelevant diagnoses in up to 53.5% of cases. Existing inference-time corrections either need counterfactual data built around the specific spurious feature or behave unpredictably. We observe that how a model's preference between answers evolves across layers, its prediction trajectory, indicates for each document whether a correction is likely to help. We propose Trajectory-Guided Steering (TGS), which decides from the trajectory alone whether to correct each prediction and in which direction, with no counterfactual data and no knowledge of the spurious feature. On 16 clinical classification tasks from i2b2, TGS fixes 17.7% of errors while breaking only 1.3% of correct predictions and stays net-positive across all tested hyperparameter settings; all baselines are net-negative or near zero. The same corrections reduce reliance on the spurious feature: on the most affected diseases, gender-swap inconsistency falls sharply (hypertension 53.5% -> 38.0%, diabetes 35.1% -> 21.5%). Applying the correction to every document instead raises breaks on correct positive predictions from 0.1% to 6.4%, so per-document selection is essential. TGS also transfers to three model families and three other clinical and biomedical datasets.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.22618 [cs.CL] |
| (or arXiv:2610.22618v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.22618 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Fardin Ahsan Sakib [view email]
[v1]
Thu, 15 Oct 2026 20:22:04 UTC (121 KB)
来源：arXiv cs.CL · arxiv.org