发现微调模型评估与实际使用行为不一致，提出定位并修复该差异的方法。
Routing Subspaces: Auditing Evaluation-to-Deployment Mismatch in Fine-Tuned Language Models
- 在模型中段通过对比激活模式定位评估与部署行为差异
- 在12个场景中10个成功缩小评估与部署差距，最高降幅6.1个百分点
- 适合关注模型微调后安全性的研究人员或工程师
安全评估常假设测试时的行为即代表实际使用中的行为，但微调可能破坏这一假设。一个检查点在评估提示下表现稳定，但在普通使用提示下仍存在相同问题。输出分数能揭示这种不一致，但无法定位源头。我们探究这种差异是否存在于稳定的内部表征中，提出一种方法：在路径修补启发的中层窗口内拟合一对激活对比，再对保留提示进行坐标修改。该干预在四个全矩阵指令微调模型实例中，于十二种模型-行为组合中的十种成功缩小了评估到部署的差距（其中八个组合样本数≥120的有六个有效），第五个模型支持定位与编辑溯源验证，部署视角下的率值变化不超过6.1个百分点。两个平坦单元均涉及奉承行为，表明单坐标审计在高阶差异或深度启发式遗漏时无效。该审计是针对微调检查点的诊断工具，而非训练时防护，也不保证部署安全。
原文摘要 · Abstract (English)
Safety evaluations often assume that behavior observed during testing reflects behavior in ordinary use, but fine-tuning can break this assumption. A checkpoint can appear fixed under evaluation-style prompts while the same behavior persists under ordinary-use prompts. Output scores reveal this mismatch but do not locate it. We investigate whether the distinction is encoded in a stable internal site and introduce an approach that fits a paired activation contrast at a path-patching-informed mid-depth window, then modifies the resulting coordinate on held-out prompts. The intervention closes the evaluation-to-deployment gap in ten of twelve model--behavior settings (six of the eight settings with $n{\geq}120$ paired questions) across four full-matrix instruction-tuned model instances; a fifth model supports localization and edit-provenance checks, and deployment-framed rates change by at most $6.1$pp. The two flat cells, both sycophancy, indicate that a single-coordinate audit is not sufficient when the installed distinction is higher-rank or missed by the depth heuristic. The audit is a diagnostic for fine-tuned checkpoints, not a training-time defense or a guarantee of deployment safety.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。