Towards a Mechanistic Understanding of Propositional Logical Reasoning in Large Language Models
cs.AI, cs.LG
Submitted: 2026-01-07
Updated: 2026-09-12
Comments: Accepted to EMNLP 2026 (Main Conference). 30 pages
Code: https://github.com/TransformerLensOrg/TransformerLens
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Decomposing The Dark Matter of Sparse Autoencoders
- How do Language Models Bind Entities in Context?
- Does Localization Inform Editing? Surprising Differences in Causality-Based Localization vs. Knowledge Editing in Language Models
- How to use and interpret activation patching
- Empowering LLMs with Logical Reasoning: A Comprehensive Survey
- A Implies B: Circuit Analysis in LLMs for Propositional Logical Reasoning
- Faith and Fate: Limits of Transformers on Compositionality
- Language Models Use Trigonometry to Do Addition
- Reasoning Circuits in Language Models: A Mechanistic Interpretation of Syllogistic Inference
- Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task
- Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models
- Locating and Editing Factual Associations in GPT
- Function Vectors in Large Language Models
- Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias
- Progress measures for grokking via mechanistic interpretability
- In-context Learning and Induction Heads
- Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small
- OpenAI o1 System Card
- Language Models can Evaluate Themselves via Probability Discrepancy
- How do Transformers Learn Implicit Reasoning?
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
- 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
- VSAL: A Vision Solver with Adaptive Layouts for Graph Property Detection