Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets

• URL: http://arxiv.org/abs/2506.05346v1
• Date: Thu, 05 Jun 2025 17:59:55 GMT
• Title: Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets
• Authors: Lei Hsiung, Tianyu Pang, Yung-Chen Tang, Linyue Song, Tsung-Yi Ho, Pin-Yu Chen, Yaoqing Yang,
• Abstract summary: 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%.
• Score: 64.96967819446553
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent advancements in large language models (LLMs) have underscored their vulnerability to safety alignment jailbreaks, particularly when subjected to downstream fine-tuning. However, existing mitigation strategies primarily focus on reactively addressing jailbreak incidents after safety guardrails have been compromised, removing harmful gradients during fine-tuning, or continuously reinforcing safety alignment throughout fine-tuning. As such, they tend to overlook a critical upstream factor: the role of the original safety-alignment data. This paper therefore investigates the degradation of safety guardrails through the lens of representation similarity between upstream alignment datasets and downstream fine-tuning tasks. Our experiments demonstrate that high similarity between these datasets significantly weakens safety guardrails, making models more susceptible to jailbreaks. Conversely, low similarity between these two types of datasets yields substantially more robust models and thus reduces harmfulness score by up to 10.33%. By highlighting the importance of upstream dataset design in the building of durable safety guardrails and reducing real-world vulnerability to jailbreak attacks, these findings offer actionable insights for fine-tuning service providers.

Related papers

• TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs [7.125400292079228]
  Large Language Models (LLMs) fine-tuned to align with human values often exhibit alignment drift.<n>While prior work has behaviorally characterized alignment failure, little is known about the training-time belief sources underlying these failures.<n>We introduce TraceAlign, a unified framework for tracing unsafe completions back to their root causes in the model's training corpus.
  arXiv  Detail & Related papers  (2025-08-04T05:03:35Z)
• Attention Slipping: A Mechanistic Understanding of Jailbreak Attacks and Defenses in LLMs [61.916827858666906]
  We reveal a universal phenomenon that occurs during jailbreak attacks: Attention Slipping.<n>We show Attention Slipping is consistent across various jailbreak methods, including gradient-based token replacement, prompt-level template refinement, and in-context learning.<n>We propose Attention Sharpening, a new defense that directly counters Attention Slipping by sharpening the attention score distribution using temperature scaling.
  arXiv  Detail & Related papers  (2025-07-06T12:19:04Z)
• Fine-Tuning Lowers Safety and Disrupts Evaluation Consistency [17.57889200051214]
  Fine-tuning a general-purpose large language model (LLM) for a specific domain or task has become a routine procedure for ordinary users.<n>We consider this to be a critical failure mode of LLMs due to the widespread uptake of fine-tuning, combined with the benign nature of the "attack"<n>Our experiments expose surprising variance in the results of the safety evaluation, even when seemingly inconsequential changes are made to the fine-tuning setup.
  arXiv  Detail & Related papers  (2025-06-20T17:57:12Z)
• AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin [38.577959886489076]
  Large language models (LLMs) are vulnerable to safety risks during fine-tuning.<n>We propose a methodology for safety fine-tuning called AsFT (Anchoring Safety in Fine-Tuning)
  arXiv  Detail & Related papers  (2025-06-10T05:59:48Z)
• SafeCOMM: What about Safety Alignment in Fine-Tuned Telecom Large Language Models? [74.5407418382515]
  Fine-tuning large language models (LLMs) for telecom tasks and datasets is a common practice to adapt general-purpose models to the telecom domain.<n>Recent research has shown that even benign fine-tuning can degrade the safety alignment of LLMs, causing them to respond to harmful or unethical user queries.
  arXiv  Detail & Related papers  (2025-05-29T13:31:51Z)
• 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)
• STAIR: Improving Safety Alignment with Introspective Reasoning [44.780098674618614]
  We propose STAIR, a framework that integrates SafeTy Alignment with Itrospective Reasoning.<n>We show that STAIR effectively mitigates harmful outputs while better preserving helpfulness, compared to instinctive alignment strategies.<n>With test-time scaling, STAIR achieves a safety performance comparable to Claude-3.5 against popular jailbreak attacks.
  arXiv  Detail & Related papers  (2025-02-04T15:02:55Z)
• Layer-Level Self-Exposure and Patch: Affirmative Token Mitigation for Jailbreak Attack Defense [55.77152277982117]
  We introduce Layer-AdvPatcher, a methodology designed to defend against jailbreak attacks.<n>We use an unlearning strategy to patch specific layers within large language models through self-augmented datasets.<n>Our framework reduces the harmfulness and attack success rate of jailbreak attacks.
  arXiv  Detail & Related papers  (2025-01-05T19:06:03Z)
• Safety Layers in Aligned Large Language Models: The Key to LLM Security [43.805905164456846]
  Internal parameters in aligned LLMs can be vulnerable to security degradation when subjected to fine-tuning attacks.<n>Our work uncovers the mechanism behind security in aligned LLMs at the parameter level, identifying a small set of contiguous layers in the middle of the model.<n>We propose a novel fine-tuning approach, Safely Partial- Fine-Tuning (SPPFT), that fixes the gradient of the safety layers during fine-tuning to address the security degradation.
  arXiv  Detail & Related papers  (2024-08-30T04:35:59Z)
• Mitigating Fine-tuning based Jailbreak Attack with Backdoor Enhanced Safety Alignment [56.2017039028998]
  Fine-tuning of Language-Model-as-a-Service (LM) introduces new threats, particularly against the Fine-tuning based Jailbreak Attack (FJAttack) We propose the Backdoor Enhanced Safety Alignment method inspired by an analogy with the concept of backdoor attacks. Our comprehensive experiments demonstrate that through the Backdoor Enhanced Safety Alignment with adding as few as 11 safety examples, the maliciously finetuned LLMs will achieve similar safety performance as the original aligned models without harming the benign performance.
  arXiv  Detail & Related papers  (2024-02-22T21:05:18Z)
• Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To! [88.90694413503614]
  We find that the safety alignment of LLMs can be compromised by fine-tuning. We jailbreak GPT-3.5 Turbo's safety guardrails by fine-tuning it on only 10 such examples. We advocate for further research efforts toward reinforcing safety protocols for the custom fine-tuning of aligned LLMs.
  arXiv  Detail & Related papers  (2023-10-05T17:12:17Z)
• Refiner: Data Refining against Gradient Leakage Attacks in Federated Learning [28.76786159247595]
  gradient leakage attacks exploit clients' uploaded gradients to reconstruct their sensitive data. In this paper, we explore a novel defensive paradigm that departs from conventional gradient perturbation approaches. We design Refiner that jointly optimize two metrics for privacy protection and performance maintenance.
  arXiv  Detail & Related papers  (2022-12-05T05:36:15Z)

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.