Causal Probing for Internal Visual Representations in Multimodal Large Language Models

ArXi:2605.05593v1 Announce Type: new Despite the remarkable success of Multimodal Large Language Models (MLLMs) across diverse tasks, the internal mechanisms governing how they encode and ground distinct visual concepts remain poorly understood. To bridge this gap, we propose a causal framework based on activation steering to actively probe and manipulate internal visual representations.