{"ID":2886702,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.02079","arxiv_id":"2508.02079","title":"AlignGuard-LoRA: Alignment-Preserving Fine-Tuning via Fisher-Guided Decomposition and Riemannian-Geodesic Collision Regularization","abstract":"Low-rank adaptation (LoRA) has become a standard tool for efficiently fine-tuning large language models (LLMs). Yet, even minor LoRA updates can induce alignment drift, weakening safety and behavioral constraints through entangled parameter changes. To address this, we propose AlignGuard-LoRA (AGL), a principled framework for preserving alignment during finetuning. AGL introduces several key components: a primary task loss for supervision, Fisher Information Matrix-based regularization to restrict updates in alignment-sensitive subspaces, and task-specific regularization to stabilize the integration of new knowledge. We further introduce collision-aware regularization, blending Riemannian overlap -- which penalizes coordinate-wise interference -- and geodesic separation -- which encourages disjoint update geometry. We curate DriftCaps, a targeted diagnostic benchmark of safe and unsafe prompts designed to quantify alignment drift and safety degradation. Empirical evaluations show that AGL mitigates alignment drift by up to 50% on safety-critical benchmarks without degrading downstream task performance. Comprehensive ablation confirms that each component contributes distinctly to preserving latent safety behaviors. Finally, we derive and validate a scaling law for catastrophic forgetting, revealing that AGL flattens post-finetuning loss escalation while preserving adaptation dynamics. AGL is a structurally grounded refinement of LoRA, ensuring alignment preservation with minimal trade-offs. To encourage further exploration and development, we open-source our implementation.","short_abstract":"Low-rank adaptation (LoRA) has become a standard tool for efficiently fine-tuning large language models (LLMs). Yet, even minor LoRA updates can induce alignment drift, weakening safety and behavioral constraints through entangled parameter changes. To address this, we propose AlignGuard-LoRA (AGL), a principled framew...","url_abs":"https://arxiv.org/abs/2508.02079","url_pdf":"https://arxiv.org/pdf/2508.02079v1","authors":"[\"Amitava Das\",\"Abhilekh Borah\",\"Vinija Jain\",\"Aman Chadha\"]","published":"2025-08-04T05:45:24Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\"]","methods":"[\"Large Language Model\",\"Language Model\",\"LoRA\"]","has_code":false}