Ravil Mussabayev
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
NLP×1AI×1
Frequent co-authors
Research Timeline
2026
MLLM-Microscope: Unlocking Hidden Structure Within Multimodal Large Language Models
The paper introduces MLLM-Microscope, a system that analyzes the internal structure of multimodal large language models (MLLMs), finding that modality fusion significantly impacts the linearity and dimensionality of token embeddings.
Highlighted terms show continued research focus across papers