Don't Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915240305754112 |
|---|---|
| author | Jiao, Pengkun Zhu, Bin Chen, Jingjing Ngo, Chong-Wah Jiang, Yu-Gang |
| author_facet | Jiao, Pengkun Zhu, Bin Chen, Jingjing Ngo, Chong-Wah Jiang, Yu-Gang |
| contents | Large Multimodal Models (LMMs) have demonstrated remarkable capabilities across a wide range of tasks. However, their vulnerability to user gaslighting-the deliberate use of misleading or contradictory inputs-raises critical concerns about their reliability in real-world applications. In this paper, we address the novel and challenging issue of mitigating the negative impact of negation-based gaslighting on LMMs, where deceptive user statements lead to significant drops in model accuracy. Specifically, we introduce GasEraser, a training-free approach that reallocates attention weights from misleading textual tokens to semantically salient visual regions. By suppressing the influence of "attention sink" tokens and enhancing focus on visually grounded cues, GasEraser significantly improves LMM robustness without requiring retraining or additional supervision. Extensive experimental results demonstrate that GasEraser is effective across several leading open-source LMMs on the GaslightingBench. Notably, for LLaVA-v1.5-7B, GasEraser reduces the misguidance rate by 48.2%, demonstrating its potential for more trustworthy LMMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_09456 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Don't Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs Jiao, Pengkun Zhu, Bin Chen, Jingjing Ngo, Chong-Wah Jiang, Yu-Gang Artificial Intelligence Computer Vision and Pattern Recognition Large Multimodal Models (LMMs) have demonstrated remarkable capabilities across a wide range of tasks. However, their vulnerability to user gaslighting-the deliberate use of misleading or contradictory inputs-raises critical concerns about their reliability in real-world applications. In this paper, we address the novel and challenging issue of mitigating the negative impact of negation-based gaslighting on LMMs, where deceptive user statements lead to significant drops in model accuracy. Specifically, we introduce GasEraser, a training-free approach that reallocates attention weights from misleading textual tokens to semantically salient visual regions. By suppressing the influence of "attention sink" tokens and enhancing focus on visually grounded cues, GasEraser significantly improves LMM robustness without requiring retraining or additional supervision. Extensive experimental results demonstrate that GasEraser is effective across several leading open-source LMMs on the GaslightingBench. Notably, for LLaVA-v1.5-7B, GasEraser reduces the misguidance rate by 48.2%, demonstrating its potential for more trustworthy LMMs. |
| title | Don't Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs |
| topic | Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2504.09456 |