Search papers, labs, and topics across Lattice.
This paper investigates how MLLMs perform image segmentation by probing representations at different stages of the pipeline: vision encoder, adapter, and LLM. Attention knockout and bidirectional attention analyses reveal that the adapter causes a drop in segmentation performance, which is then recovered in the LLM layers through attention-mediated refinement. The recovery process is initially limited by causal attention, but bidirectional attention among image tokens can alleviate this limitation.
MLLMs' image segmentation prowess isn't a given: a critical adapter layer actually *hurts* performance, with the LLM having to recover via attention-mediated refinement.
Multimodal Large Language Models (MLLMs) are increasingly applied to pixel-level vision tasks, yet their intrinsic capacity for spatial understanding remains poorly understood. We investigate segmentation capacity through a layerwise linear probing evaluation across the entire MLLM pipeline: vision encoder, adapter, and LLM. We further conduct an intervention based attention knockout analysis to test whether cross-token attention progressively refines visual representations, and an evaluation of bidirectional attention among image tokens on spatial consistency. Our analysis reveals that the adapter introduces a segmentation representation drop-off, but LLM layers progressively recover through attention-mediated refinement, where correctly classified tokens steer misclassified neighbors toward the correct label. At early image token positions, this recovery is bounded by causal attention, which bidirectional attention among image tokens alleviates. These findings provide a mechanistic account of how MLLMs process visual information for segmentation, informing the design of future segmentation-capable models.
