No article text was recorded for this version.
post-hoc interpretability overhead
Version 1 of 127 papers27 conceptsfrozen 2026-10-04
Papers in this version
- Blog: Beyond Interpretability: The Gains of Feature Monosemanticity on Model Robustness
- Blog: Group-SAE: Efficient Training of Sparse Autoencoders for Large Language Models via Layer Groups
- Blog: Atlas-Alignment: Making Interpretability Transferable Across Language Models
- Blog: Learning Retrieval Models with Sparse Autoencoders
- Blog: Mixture of Experts Made Intrinsically Interpretable
- Blog: Transferring Linear Features Across Language Models With Model Stitching
- Blog: Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
- Blog: Learning by Interpreting
- Blog: Weight-based Decomposition: A Case for Bilinear MLPs
- Blog: Inference-Time Decomposition of Activations (ITDA): A Scalable Approach to Interpreting Large Language Models
- Blog: Semantic Convergence: Investigating Shared Representations Across Scaled LLMs
- Blog: Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering
- Blog: Singular Vectors of Attention Heads Align with Features
- Blog: Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models
- Blog: SAE as a Crystal Ball: Interpretable Features Predict Cross-domain Transferability of LLMs without Training
- Blog: Boosting-based Parse Reranking with Subtree Features
- Blog: Interpreting Genomic Language Models using Sparse Autoencoders
- Blog: Causal Interpretation of Neural Network Computations with Contribution Decomposition
- Blog: Beyond Redundancy: Diverse and Specialized Multi-Expert Sparse Autoencoder
- Blog: What makes your model a low-empathy or warmth person: Exploring the Origins of Personality in LLMs
- Blog: Automatically Identifying Local and Global Circuits with Linear Computation Graphs
- Blog: SVD as a Fast Interpretability Method for Transformers
- Blog: The Knowledge Microscope: Features as Better Analytical Lenses than Neurons
- Blog: Exploring Accurate and Transparent Domain Adaptation in Predictive Healthcare via Concept-Grounded Orthogonal Inference
- Blog: Concept-Guided Interpretability via Neural Chunking
- Blog: Transformer visualization via dictionary learning: contextualized embedding as a linear superposition of transformer factors
- Blog: Self-Regularization with Sparse Autoencoders for Controllable LLM-based Classification
History & citation
Current · Permanent link
- This direction split from Monosemantic Feature Discovery in Polysemantic Neural Activations (74% overlap)
Cite this direction
Latest version (a living page):
@misc{lacuna_post_hoc_interpretability_overhead_d0cdb94b,
title = {post-hoc interpretability overhead},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/post-hoc-interpretability-overhead/rdr_d0cdb94b3174463197cf572e3248201e},
note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}This exact version (version 1, frozen text):
@misc{lacuna_post_hoc_interpretability_overhead_d0cdb94b_v1,
title = {post-hoc interpretability overhead},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/post-hoc-interpretability-overhead/rdr_d0cdb94b3174463197cf572e3248201e/v/1},
note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}