基于特征空间监测监督微调中的涌现性错位

View PDF HTML (experimental)

Abstract:Emergent misalignment (EM) occurs when narrow finetuning induces dangerous behavior outside the finetuning task. Detecting this shift through repeated behavioral evaluation is costly, motivating our checkpoint-level monitoring from internal representations. We define a fixed coordinate system from seven alignment-relevant activation directions and use it to track representational drift during LoRA finetuning of four open-source 7-9B language models. Finetuning drift in this space exhibits a dominant axis that explains 78.6% of variance and remains stable across datasets, extraction choices, and parameter-update capacities. Across 468 checkpoints from three EM-relevant held-out datasets, the resulting monitors attain 1.8% FNR, 2.0% FPR, and 0.989 AUROC, outperforming semantic, random, PCA, and SAE feature baselines. On a fourth dataset, a matched benign-dangerous control shows that substantial representational drift can also occur under benign finetuning, while changes across the 7D profile still distinguish dangerous from benign runs. Stress tests across two 14B models, full finetuning, longer training horizons, and misaligned starting states show that the signal can persist across shifts in training configuration, while reliable deployment may require recalibration.

Submission history

From: Huy Nghiem [view email]
[v1] Sun, 31 May 2026 04:28:21 UTC (291 KB)
[v2] Thu, 1 Oct 2026 16:52:27 UTC (302 KB)