Decodable but Not Actionable: Auditing the Representation-Behavior Gap in Frozen LLMs for Geometric Constraints
cs.CL, cs.AI
Submitted: 2026-08-18
Updated: 2026-10-27
Comments: Revised manuscript with expanded experiments, analyses, and discussion
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- The Llama 3 Herd of Models
- Language Models Represent Space and Time
- Mistral 7B
- Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task
- Emergent Linear Representations in World Models of Self-Supervised Sequence Models
- The Linear Representation Hypothesis and the Geometry of Large Language Models
- Qwen2.5 Technical Report
- Steering Language Models With Activation Engineering
- No Training Required: Exploring Random Encoders for Sentence Classification
- Fusion 360 Gallery: A Dataset and Environment for Programmatic CAD Construction from Human Design Sequences
- AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation
- Representation Engineering: A Top-Down Approach to AI Transparency
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data