Narrow fine-tuning erodes safety alignment in vision-language agents

• URL: http://arxiv.org/abs/2602.16931v1
• Date: Wed, 18 Feb 2026 22:47:28 GMT
• Title: Narrow fine-tuning erodes safety alignment in vision-language agents
• Authors: Idhant Gulati, Shivam Raval,
• Abstract summary: Lifelong multimodal agents must continuously adapt to new tasks through post-training.<n>We demonstrate that fine-tuning aligned vision-language models on narrow-domain harmful datasets induces severe emergent misalignment.
• Score: 0.12441041004077093
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Lifelong multimodal agents must continuously adapt to new tasks through post-training, but this creates fundamental tension between acquiring capabilities and preserving safety alignment. We demonstrate that fine-tuning aligned vision-language models on narrow-domain harmful datasets induces severe emergent misalignment that generalizes broadly across unrelated tasks and modalities. Through experiments on Gemma3-4B, we show that misalignment scales monotonically with LoRA rank, and that multimodal evaluation reveals substantially higher misalignment ($70.71 \pm 1.22$ at $r=128$) than text-only evaluation ($41.19 \pm 2.51$), suggesting that unimodal safety benchmarks may underestimate alignment degradation in vision-language models. Critically, even 10\% harmful data in the training mixture induces substantial alignment degradation. Geometric analysis reveals that harmful behaviors occupy a remarkably low-dimensional subspace, with the majority of misalignment information captured in 10 principal components. To mitigate misalignment, we evaluate two strategies: benign narrow fine-tuning and activation-based steering. While both approaches substantially reduce misalignment, neither completely removes the learned harmful behaviors. Our findings highlight the need for robust continual learning frameworks, as current post-training paradigms may not sufficiently preserve alignment in post-deployment settings.

Related papers

• Position: General Alignment Has Hit a Ceiling; Edge Alignment Must Be Taken Seriously [51.03213216886717]
  We take the position that the dominant paradigm of General Alignment reaches a structural ceiling in settings with conflicting values.<n>We introduce Edge Alignment as a distinct approach in which systems preserve multi dimensional value structure.
  arXiv  Detail & Related papers  (2026-02-23T16:51:43Z)
• OmniVL-Guard: Towards Unified Vision-Language Forgery Detection and Grounding via Balanced RL [63.388513841293616]
  Existing forgery detection methods fail to handle the interleaved text, images, and videos prevalent in real-world misinformation.<n>To bridge this gap, this paper targets to develop a unified framework for omnibus vision-language forgery detection and grounding.<n>We propose textbf OmniVL-Guard, a balanced reinforcement learning framework for omnibus vision-language forgery detection and grounding.
  arXiv  Detail & Related papers  (2026-02-11T09:41:36Z)
• Understanding Degradation with Vision Language Model [56.09241449206817]
  Understanding visual degradations is a critical yet challenging problem in computer vision.<n>We introduce DU-VLM, a multimodal chain-of-thought model trained with supervised fine-tuning and reinforcement learning.<n>We also introduce textbfDU-110k, a large-scale dataset comprising 110,000 clean-degraded pairs with grounded physical annotations.
  arXiv  Detail & Related papers  (2026-02-04T13:51:15Z)
• Character as a Latent Variable in Large Language Models: A Mechanistic Account of Emergent Misalignment and Conditional Safety Failures [70.48661957773449]
  Emergent Misalignment refers to a failure mode in which fine-tuning large language models on narrowly scoped data induces broadly misaligned behavior.<n>Across multiple domains and model families, we find that fine-tuning models on data exhibiting specific character-level dispositions induces substantially stronger and more transferable misalignment than incorrect-advice fine-tuning.
  arXiv  Detail & Related papers  (2026-01-30T15:28:42Z)
• The Realignment Problem: When Right becomes Wrong in LLMs [6.8304813545377]
  The alignment of Large Language Models with human values is central to their safe deployment, yet current models fail to keep pace with evolving norms and policies.<n>Existing unlearning methods act as blunt instruments that erode utility rather than enable precise policy updates.<n>We introduce TRACE, a framework for principled unlearning that reconceives realignment as a programmatic policy problem.
  arXiv  Detail & Related papers  (2025-11-04T14:52:58Z)
• DUAL-Bench: Measuring Over-Refusal and Robustness in Vision-Language Models [59.45605332033458]
  Safety mechanisms can backfire, causing over-refusal, where models decline benign requests out of excessive caution.<n>No existing benchmark has systematically addressed over-refusal in the visual modality.<n>This setting introduces unique challenges, such as dual-use cases where an instruction is harmless, but the accompanying image contains harmful content.
  arXiv  Detail & Related papers  (2025-10-12T23:21:34Z)
• Re-Emergent Misalignment: How Narrow Fine-Tuning Erodes Safety Alignment in LLMs [0.0]
  We show that fine tuning on insecure code induces internal changes that oppose alignment.<n>We identify a shared latent dimension in the model's activation space that governs alignment behavior.
  arXiv  Detail & Related papers  (2025-07-04T15:36:58Z)
• Probing the Robustness of Large Language Models Safety to Latent Perturbations [30.16804362984161]
  Safety alignment is a key requirement for building reliable Artificial General Intelligence.<n>We observe that minor latent shifts can still trigger unsafe responses in aligned models.<n>We introduce Layer-wise Adversarial Patch Training(LAPT), a fine-tuning strategy that injects controlled perturbations into hidden representations during training.
  arXiv  Detail & Related papers  (2025-06-19T07:03:05Z)
• Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets [64.96967819446553]
  This paper investigates the degradation of safety guardrails through the lens of representation similarity between upstream alignment datasets and downstream fine-tuning tasks.<n>High similarity between these datasets significantly weakens safety guardrails, making models more susceptible to jailbreaks.<n>Low similarity between these two types of datasets yields substantially more robust models and thus reduces harmfulness score by up to 10.33%.
  arXiv  Detail & Related papers  (2025-06-05T17:59:55Z)
• Do We Really Need Curated Malicious Data for Safety Alignment in Multi-modal Large Language Models? [83.53005932513155]
  Multi-modal large language models (MLLMs) have made significant progress, yet their safety alignment remains limited.<n>We propose finetuning MLLMs on a small set of benign instruct-following data with responses replaced by simple, clear rejection sentences.
  arXiv  Detail & Related papers  (2025-04-14T09:03:51Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.