Hierarchical Safety Realignment: Lightweight Restoration of Safety in Pruned Large Vision-Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Recent work has shown that pruning can reduce model performance, but it can also lead to degradation in safety performance. |
| Approach: | They propose a hierarchical safety realignment approach to prune large vision-Language Models . they quantify contribution of each attention head to safety and restore neurons . |
| Outcome: | The proposed approach achieves significant safety improvements in LVLMs pruned post pruning. |
Similar Papers
SafeMERGE: Preserving Safety Alignment in Fine-Tuned Large Language Models via Selective Layer-Wise Model Merging (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent studies show that fine-tuning can erode safety alignment, causing LLMs to respond to harmful or unethical prompts. |
| Approach: | They propose a lightweight framework that restores safety while maintaining downstream performance. |
| Outcome: | The proposed framework reduces harmful outputs compared to other defenses, with negligible impact on utility. |
Seeing No Evil: Blinding Large Vision-Language Models to Safety Instructions via Adversarial Attention Hijacking (2026.acl-long)
Copied to clipboard
| Challenge: | Existing attacks optimize image perturbations to maximize harmful output likelihood, but suffer from slow convergence due to gradient conflict between adversarial objectives and the model’s safety-retrieval mechanism. |
| Approach: | They propose a push-pull approach which suppresses attention to system-prompt tokens and anchors generation on adversarial image features to avoid collisions. |
| Outcome: | The proposed approach reduces gradient conflict by 45% and achieves 94.4% attack success rate on Qwen-VL (vs. 68.8% baseline) with 40% fewer iterations. |
Reducing Token Redundancy in LVLMs: A Systematic Review of Token Pruning Methods (2026.acl-long)
Copied to clipboard
| Challenge: | Large Vision-Language Models (LVLMs) excel at visual understanding but face severe computational bottlenecks when processing high-resolution images and long videos due to massive visual token counts. |
| Approach: | They propose a taxonomy categorizing methods into vision-side, LLM-side and hybrid paradigms and analyze token selection mechanisms and pruning strategy. |
| Outcome: | The proposed method selectively removes less informative tokens while maintaining performance. |
RankAdaptor: Hierarchical Rank Allocation for Efficient Fine-Tuning Pruned LLMs via Performance Model (2025.findings-naacl)
Copied to clipboard
| Challenge: | Current compression techniques entail structural pruning and a recovery phase that leverages the Low-Rank Adaptation algorithm. |
| Approach: | They propose a hierarchical rank allocation method that enables efficient fine-tuning of pruned LLMs according to layerwise specific recovery requirements. |
| Outcome: | The proposed algorithm outperforms state-of-the-art methods across pruning settings and LLM architectures with improvements ranging from 0.7% to 5.5%. |
LVPruning: An Effective yet Simple Language-Guided Vision Token Pruning Approach for Multi-modal Large Language Models (2025.findings-naacl)
Copied to clipboard
| Challenge: | Multi-modal Large Language Models (MLLMs) incur significant computational overhead due to the large number of vision tokens processed, limiting their practicality in resource-constrained environments. |
| Approach: | They propose a language-guided vision token pruning method that can be integrated into existing MLLMs with minimal architectural changes. |
| Outcome: | The proposed method reduces vision tokens by 90% and preserves model performance. |
LSSF: Safety Alignment for Large Language Models through Low-Rank Safety Subspace Fusion (2025.acl-long)
Copied to clipboard
| Challenge: | Existing safety alignment methods rely on fine-tuning, which inadvertently leads to the increased complexity and computational resources required. |
| Approach: | They propose a safety re-alignment framework with Low-Rank Safety Subspace Fusison that exploits low-rank safety characteristics of LLMs by constructing a low-ranked projection matrix to extract the principal components of safety vectors. |
| Outcome: | The proposed method exploits low-rank safety subspace of the LLMs and is stable during fine-tuning process and is isolated from the model’s general capabilities. |
Separate the Wheat from the Chaff: A Post-Hoc Approach to Safety Re-Alignment for Fine-Tuned Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Large language models achieve effective safety alignment at the time of release, but fine-tuning often compromises safety mechanisms. |
| Approach: | They propose a method that performs safety realignment for large language models . they identify unsafe delta parameters from the fine-tuned models and recalibrate the retained parameters . |
| Outcome: | The proposed method improves safety performance on safety benchmarks and jailbreak attacks while maintaining their performance on downstream tasks. |
Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Fine-tuning large language models for downstream tasks often leads to catastrophic forgetting, notably degrading the safety of original alignments. |
| Approach: | They propose to merge the weights of pre- and post-fine-tuned models to improve safety while enhancing performance. |
| Outcome: | Experiments across different downstream tasks and models validate the method’s practicality and effectiveness. |
FLASH: Focused Layer Attention Sink Hijacking (2026.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) remain vulnerable to jailbreaking attacks despite advances in safety alignment . |
| Approach: | They propose a new diagnostic auditing framework that dismantles the model's internal safety anchor by precisely scaling attention scores in these vulnerable layers. |
| Outcome: | The proposed framework achieves a state-of-the-art Attack Success Rate of over 77% with an unprecedented efficiency of 1.53 queries on average. |
BlockPruner: Fine-grained Pruning for Large Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) have significant computational and memory costs associated with training and inference. |
| Approach: | They propose a training-free structured pruning approach that targets redundancies in MHA and MLP blocks. |
| Outcome: | The proposed pruning approach achieves more granular and effective pruning compared to state-of-the-art pruning methods. |