Don't Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs

Don't Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs

13 April 2025

ArXiv (abs)PDF HTML

Papers citing "Don't Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs"

9 / 9 papers shown

Title
See What You Are Told: Visual Attention Sink in Large Multimodal Models Seil Kang Jinyeong Kim Junhyeok Kim Seong Jae Hwang VLM 159 10 0 05 Mar 2025
Calling a Spade a Heart: Gaslighting Multimodal Large Language Models via Negation Bin Zhu Hui yan Qi Yinxuan Gui Jingjing Chen Chong-Wah Ngo Ee-Peng Lim 410 2 0 31 Jan 2025
Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling Xiaokang Chen Zhiyu Wu Xingchao Liu Zizheng Pan Wen Liu Zhenda Xie X. Yu Chong Ruan AI4TS 152 160 0 29 Jan 2025
Language models are not naysayers: An analysis of language models on negation benchmarks Thinh Hung Truong Timothy Baldwin Karin Verspoor Trevor Cohn 116 60 0 14 Jun 2023
When and why vision-language models behave like bags-of-words, and what to do about it? Mert Yuksekgonul Federico Bianchi Pratyusha Kalluri Dan Jurafsky James Zou VLM CoGe 104 394 0 04 Oct 2022
PaLI: A Jointly-Scaled Multilingual Language-Image Model Xi Chen Tianlin Li Soravit Changpinyo A. Piergiovanni Piotr Padlewski ... Andreas Steiner A. Angelova Xiaohua Zhai N. Houlsby Radu Soricut MLLM VLM 137 736 0 14 Sep 2022
Emerging Properties in Self-Supervised Vision Transformers Mathilde Caron Hugo Touvron Ishan Misra Hervé Jégou Julien Mairal Piotr Bojanowski Armand Joulin 735 6,139 0 29 Apr 2021
Negated and Misprimed Probes for Pretrained Language Models: Birds Can Talk, But Cannot Fly Nora Kassner Hinrich Schütze 76 324 0 08 Nov 2019
Attention Is All You Need Ashish Vaswani Noam M. Shazeer Niki Parmar Jakob Uszkoreit Llion Jones Aidan Gomez Lukasz Kaiser Illia Polosukhin 3DV 808 132,725 0 12 Jun 2017

We use cookies and other tracking technologies to improve your browsing experience on our website, to show you personalized content and targeted ads, to analyze our website traffic, and to understand where our visitors are coming from. See our policy.