mlsanityData to 14 Oct 2026

HyperDAS: Towards Automating Mechanistic Interpretability with Hypernetworks

arXiv 2503.10894 ↗First posted 13 Mar 2025cs.CL cs.AI cs.LGcited by 9

On the map: Mechanistic interpretability → Circuits and mechanistic analysis

Abstract (arXiv)

Mechanistic interpretability has made great strides in identifying neural network features (e.g., directions in hidden activation space) that mediate concepts(e.g., the birth year of a person) and enable predictable manipulation. Distributed alignment search (DAS) leverages supervision from counterfactual data to learn concept features within hidden states, but DAS assumes we can afford to conduct a brute force search over potential feature locations. To address this, we present HyperDAS, a transformer-based hypernetwork architecture that (1) automatically locates the token-positions of the residual stream that a concept is realized in and (2) constructs features of those residual stream vectors for the concept. In experiments with Llama3-8B, HyperDAS achieves state-of-the-art performance on the RAVEL benchmark for disentangling concepts in hidden states. In addition, we review the design decisions we made to mitigate the concern that HyperDAS (like all powerful interpretabilty methods) might inject new information into the target model rather than faithfully interpreting it.

Key facts. Cited by 9 papers in the map (data through 14 Oct 2026). Placed in Circuits and mechanistic analysis, one of 13 directions in Mechanistic interpretability. Builds on 23 of its 32 references that are in the map. First posted on arXiv 13 Mar 2025.

Builds on

23 of 32 references.

Closest ancestors · 2

Same direction, newest first

  1. 2023Towards Automated Circuit Discovery for Mechanistic Interpretability28 Apr
  2. 2023Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretability11 Jan

Other references · 21

  1. 2022Locating and Editing Factual Associations in GPTcited by 2,691
  2. 2023Sparse Autoencoders Find Highly Interpretable Features in Language Modelscited by 1,420
  3. 2022Mass-Editing Memory in a Transformercited by 1,001
  4. 2023Dissecting Recall of Factual Associations in Auto-Regressive Language Modelscited by 528
  5. 2024Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Modelscited by 435
  6. 2024Patchscopes: A Unifying Framework for Inspecting Hidden Representations of Language Modelscited by 235
  7. 2023Finding Alignments Between Interpretable Causal Variables and Distributed Neural Representationscited by 209
  8. 2020Compositional Explanations of Neuronscited by 172
  9. 2023Inspecting and Editing Knowledge Representations in Language Modelscited by 141
  10. 2022Natural Language Descriptions of Deep Visual Featurescited by 137
  11. 2021Causal Analysis of Syntactic Agreement Mechanisms in Neural Language Modelscited by 131
  12. 2023Interpretability at Scale: Identifying Causal Mechanisms in Alpacacited by 103
9 more
  1. 2024Information Flow Routes: Automatically Interpreting Language Models at Scalecited by 95
  2. 2023A Mechanistic Interpretation of Arithmetic Reasoning in Language Models using Causal Mediation Analysiscited by 89
  3. 2024RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representationscited by 79
  4. 2024Identifying Functionally Important Features with End-to-End Sparse Dictionary Learningcited by 67
  5. 2024A Multimodal Automated Interpretability Agentcited by 63
  6. 2024Recurrent Neural Networks Learn to Store and Generate Sequences using Non-Linear Representationscited by 47
  7. 2023Rigorously Assessing Natural Language Explanations of Neuronscited by 45
  8. 2023FIND: A Function Description Benchmark for Evaluating Interpretability Methodscited by 32
  9. 2024Automatic Discovery of Visual Circuitscited by 13

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 4; 2026: 5.

9 papers.

  1. 2025The Non-Linear Representation Dilemma: Is Causal Abstraction Enough for Mechanistic Interpretability?cited by 37
  2. 2025HyperSteer: Activation Steering at Scale with Hypernetworkscited by 23
  3. 2025Predictive Concept Decoders: Training Scalable End-to-End Interpretability Assistantscited by 12
  4. 2026From Directions to Regions: Decomposing Activations in Language Models via Local Geometrycited by 12
  5. 2025Internal Causal Mechanisms Robustly Predict Language Model Out-of-Distribution Behaviorscited by 11
  6. 2026Formal Mechanistic Interpretability: Automated Circuit Discovery with Provable Guaranteescited by 11
  7. 2026Mechanistic Indicators of Steering Effectiveness in Large Language Modelscited by 0
  8. 2026PLOT: Progressive Localization via Optimal Transport in Neural Causal Abstractioncited by 0
  9. 2026Capacity, Responsiveness and Alignment: What Makes a Latent Structure Actionablecited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Generating Interpretable Networks using Hypernetworks
  2. 2026Mechanistic Interpretability for Neural Networks: Circuits, Sparse Features and Symbolic Reasoning
  3. 2026Mechanistic Interpretability for Large Language Model Alignment: Progress, Challenges, and Future Directions
  4. 2026Tracking Equivalent Mechanistic Interpretations Across Neural Networks
  5. 2024A Practical Review of Mechanistic Interpretability for Transformer-Based Language Models
  6. 2025Unboxing the Black Box: Mechanistic Interpretability for Algorithmic Understanding of Neural Networks
  7. 2025Position: Mechanistic Interpretability Should Prioritize Feature Consistency in SAEs
  8. 2022Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small