No article text was recorded for this version.
causal interpretability
Version 1 of 145 papers49 conceptsfrozen 2026-10-04
Papers in this version
- Blog: Verification of the Implicit World Model in a Generative Model via Adversarial Sequences
- Blog: Emergent Stack Representations in Modeling Counter Languages Using Transformers
- Blog: CheckersGPT: Learning World Models through Language Modeling
- Blog: Do LLMs Build World Representations? Probing Through the Lens of State Abstraction
- Blog: Influence-Directed Explanations for Deep Convolutional Networks
- Blog: Causal Discovery and Inference through Next-Token Prediction
- Blog: Revisiting the Importance of Individual Units in CNNs via Ablation
- Blog: Emergent Structured Representations Support Flexible In-Context Inference in Large Language Models
- Blog: Categorize Early, Integrate Late: Divergent Processing Strategies in Automatic Speech Recognition
- Blog: Transformers Represent Belief State Geometry in their Residual Stream
- Blog: Pixology: Probing the Linguistic and Visual Capabilities of Pixel-based Language Models
- Blog: Separating Tongue from Thought: Activation Patching Reveals Language-Agnostic Concept Representations in Transformers
- Blog: Rigorously Assessing Natural Language Explanations of Neurons
- Blog: Can Interpretation Predict Behavior on Unseen Data?
- Blog: Probing for Incremental Parse States in Autoregressive Language Models
- Blog: Sparse Autoencoders Reveal Temporal Difference Learning in Large Language Models
- Blog: Understanding the Role of Individual Units in a Deep Neural Network
- Blog: Evaluating Contextualized Representations of (Spanish) Ambiguous Words: A New Lexical Resource and Empirical Analysis
- Blog: What if This Modified That? Syntactic Interventions with Counterfactual Embeddings
- Blog: Transformers Use Causal World Models in Maze-Solving Tasks
- Blog: On the relationship between visual attributes and convolutional networks
- Blog: Linear Relational Decoding of Morphology in Language Models
- Blog: From Associations to Activations: Comparing Behavioral and Hidden-State Semantic Geometry in LLMs
- Blog: Single-Position Intervention Fails: Distributed Output Templates Drive In-Context Learning
- Blog: Emergence of Separable Manifolds in Deep Language Representations
- Blog: Towards a Mechanistic Interpretation of Multi-Step Reasoning Capabilities of Language Models
- Blog: AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders
- Blog: Quiet Feature Learning in Algorithmic Tasks
- Blog: Causality $\neq$ Decodability, and Vice Versa: Lessons from Interpreting Counting ViTs
- Blog: Interpreting Emergent Planning in Model-Free Reinforcement Learning
- Blog: Virtual perturbations to assess explainability of deep-learning based cell fate predictors
- Blog: When Depth Adds Nothing
- Blog: Linearly Structured World Representations in Maze-Solving Transformers
- Blog: Using Shapley interactions to understand how models use structure
- Blog: Internal states before wait modulate reasoning patterns
- Blog: Unit Testing for Concepts in Neural Networks
- Blog: Dissociating Decodability and Causal Use in Bracket-Sequence Transformers
- Blog: Interpreting the Latent Structure of Operator Precedence in Language Models
- Blog: Evaluation beyond Task Performance: Analyzing Concepts in AlphaZero in Hex
- Blog: Causal Interpretation of Neural Network Computations with Contribution Decomposition
- Blog: Distributional reasoning in LLMs: Parallel reasoning processes in multi-hop reasoning
- Blog: Understanding Game-Playing Agents with Natural Language Annotations
- Blog: Representational Analysis of Binding in Language Models
- Blog: VISIT: Visualizing and Interpreting the Semantic Information Flow of Transformers
- Blog: Spying on Your Neighbors: Fine-grained Probing of Contextual Embeddings for Information about Surrounding Words
History & citation
Current · Permanent link
Cite this direction
Latest version (a living page):
@misc{lacuna_causal_interpretability_444da32c,
title = {causal interpretability},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/causal-interpretability/rdr_444da32ca61f4383ba9ef4fdfdc9c8e3},
note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}This exact version (version 1, frozen text):
@misc{lacuna_causal_interpretability_444da32c_v1,
title = {causal interpretability},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/causal-interpretability/rdr_444da32ca61f4383ba9ef4fdfdc9c8e3/v/1},
note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}