Through the LENS: Local Geometric Decomposition of Vision-Language Model Representations
Shalom Kachko, Raz Lapid, Margarita Vald, Almog Dubin, Moshe Sipper
cs.AI, cs.CL, cs.CV
Submitted: 2026-08-01
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Qwen2.5-VL Technical Report
- A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders
- Toy Models of Superposition
- Adam: A Method for Stochastic Optimization
- A Survey on Mechanistic Interpretability for Multi-Modal Foundation Models
- UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction
- Sparse Autoencoders Learn Monosemantic Features in Vision-Language Models
- Qwen2.5 Technical Report
- From Directions to Regions: Decomposing Activations in Language Models via Local Geometry
- Same Content, Different Answers: Cross-Modal Inconsistency in MLLMs
- Qwen3 Technical Report
- A Survey on Foundation-Model-Based Industrial Defect Detection
- Multimodal Chain-of-Thought Reasoning in Language Models
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems