Metonymic Circuits for Abstract Concept Grounding in Vision Transformers

arXiv:2610.06928 · cs.AI · Submitted 2026-10-03 · Read on arXiv

cs.AI

Submitted: 2026-10-03

Updated: 2026-10-03

Code: https://github.com/jingjingjing-ding/metonymic-circuits

Terminology

• Grounding — 1× in this paper · explained in 3 episodes
• Transformer — 1× in this paper · explained in 2 episodes

Sources

• Revisiting Referring Expression Comprehension Evaluation in the Era of Large Multimodal Models
• Shikra: Unleashing Multimodal LLM's Referential Dialogue Magic
• Sparse Autoencoders Find Highly Interpretable Features in Language Models
• DECEPTICON: How Dark Patterns Manipulate Web Agents
• DataComp: In search of the next generation of multimodal datasets
• Exploring Visual Interpretability for Contrastive Language-Image Pre-training
• Interpretable and Testable Vision Features via Sparse Autoencoders
• Qwen3 Technical Report
• The Mechanistic Emergence of Symbol Grounding in Language Models
• Circuit Tracing in Vision-Language Models: Understanding the Internal Mechanisms of Multimodal Thinking

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