Mechanistic Interpretability papers, 2023–2026

Key findings

• Mechanistic Interpretability went from 17 papers in 2023 to 138 in 2026, roughly 8.1x in three years.
• ICML publishes the most Mechanistic Interpretability work: 119 papers from 2023 to 2026, 38% of the four-venue total.
• The fastest-growing overlap is with Deep learning · large language models: papers tagged with both rose from 1 in 2023 to 23 in 2026.
• Among the 5 topics under Interpretability, it ranks #2 by 2026 paper count; the fastest grower in the group, Feature Attribution, is up 1.9x versus 8.1x here.

Representative papers

Spread across years and venues so every edition shows up

• NeurIPS 2026A Mechanistic Interpretability Study of an Astronomical Foundation Model
• ICML 2026Contribution Weights: A Geometrical Analysis of Self-Attention Transformers
• ICLR 2026Circuit Insights: Towards Interpretability Beyond Activations
• CVPR 2026CIGMA Causal Information-Gain Mechanistic Attribution of Attention Heads in Vision
• NeurIPS 2025Beyond Components: Singular Vector-Based Interpretability of Transformer Circuits
• ICML 2025Algorithm Development in Neural Networks: Insights from the Streaming Parity Task
• ICLR 2025Arithmetic Without Algorithms: Language Models Solve Math with a Bag of Heuristics
• CVPR 2025Dissecting and Mitigating Diffusion Bias via Mechanistic Interpretability
• NeurIPS 2024InterpBench: Semi-Synthetic Transformers for Evaluating Mechanistic Interpretability Techniques
• ICML 2024Mechanistic Neural Networks for Scientific Machine Learning
• ICLR 2024Fine-Tuning Enhances Existing Mechanisms: A Case Study on Entity Tracking
• CVPR 2024SI-MIL: Taming Deep MIL for Self-Interpretability in Gigapixel Histopathology
• NeurIPS 2023A Logic for Expressing Log-Precision Transformers
• ICML 2023Mechanistic Mode Connectivity
• ICLR 2023Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 Small
• NeurIPS 2026Discovering Mechanistic Models of Neural Activity: System Identification in an in Silico Zebrafish
• ICML 2026CorrSteer: Generation-Time LLM Steering via Correlated Sparse Autoencoder Features
• ICLR 2026CLUE: Conflict-guided Localization for LLM Unlearning Framework
• CVPR 2026From Weights to Concepts Data-Free Interpretability of CLIP via Singular
• NeurIPS 2025Bigram Subnetworks: Mapping to Next Tokens in Transformer Language Models

All Mechanistic Interpretability papers

74 papers, page 1 of 2

1. ICML 2026A Capacity-Based Rationale for Multi-Head Attention
2. ICML 2026A Distributional View for Visual Mechanistic Interpretability: KL-Minimal Soft-Constraint Principle
3. ICML 2026A Fourier perspective on the learning dynamics of neural networks: from sample complexities to mechanistic insights
4. ICML 2026A Narrowing Geometry in Contaminated Reasoning
5. ICML 2026Adversarial Vulnerability from Interference Between Features in Superposition
6. ICML 2026AI Engram: In Search of Memory Traces in Artificial Intelligence
7. ICML 2026All Circuits Lead to Rome: Rethinking Functional Anisotropy in Circuit and Sheaf Discovery for LLMs
8. ICML 2026Bayesian Gated Non-Negative Contrastive Learning
9. ICML 2026Beyond Fixed Biases: Decoding the Role of Reasoning Uncertainty in MLLM Modality Conflicts
10. ICML 2026BLOCK-EM: Preventing Emergent Misalignment via Latent Blocking
11. ICML 2026Breaking the Block: Preserving Data Continuity to Train Superior SAEs for Instruct Models
12. ICML 2026Certified Circuits: Stability Guarantees for Mechanistic Circuits
13. ICML 2026CircuitPrint: Mechanistic Circuit Fingerprints for Large Language Models
14. ICML 2026Contribution Weights: A Geometrical Analysis of Self-Attention Transformers
15. ICML 2026CorrSteer: Generation-Time LLM Steering via Correlated Sparse Autoencoder Features
16. ICML 2026DB-KSVD: Scalable Alternating Optimization for Disentangling High-Dimensional Embedding Spaces
17. ICML 2026Deep neural networks divide and conquer dihedral multiplication
18. ICML 2026Discovering Interpretable Algorithms by Decompiling Transformers to RASP
19. ICML 2026DLM-Scope: Mechanistic Interpretability of Diffusion Language Models via Sparse Autoencoders
20. ICML 2026Do Language Models Track Entities Across State Changes?
21. ICML 2026Do Sparse Autoencoders Identify Reasoning Features in Language Models?
22. ICML 2026Dual Mechanisms of Value Expression: Intrinsic vs. Prompted Values in Large Language Models
23. ICML 2026Endogenous Resistance to Activation Steering in Language Models
24. ICML 2026Faster Query-Key Learning Sharpens Attention in Self-Attention Models
25. ICML 2026From Directions to Regions: Decomposing Activations in Language Models via Local Geometry
26. ICML 2026Hide and Seek in Embedding Space: Geometry-based Steganography and Detection in Large Language Models
27. ICML 2026How Do Language Models Speak Languages? A Case Study on Unintended Code-Switching
28. ICML 2026How do LLMs Compute Verbal Confidence?
29. ICML 2026How Few-Shot Examples Add Up: A Causal Decomposition of Function Vectors in In-Context Learning
30. ICML 2026How Language Models Process Negation
31. ICML 2026Induction Meets Biology: Mechanisms of Repeat Detection in Protein Language Models
32. ICML 2026Interpretability Transfer from Language to Vision via Sparse Autoencoders
33. ICML 2026Interpreting Physics in Video World Models
34. ICML 2026Language as a Wave Phenomenon: Semantic Phase Locking and Interference in Neural Networks
35. ICML 2026Language Model Circuits Are Sparse in the Neuron Basis
36. ICML 2026Learning Adaptive Perturbation-Conditioned Contexts for Robust Transcriptional Response Prediction
37. ICML 2026Learning Protein Structure-Function Relationships through Knowledge-guided Representation Decomposition
38. ICML 2026Learning Self-Interpretation from Interpretability Artifacts: Training Lightweight Adapters on Vector-Label Pairs
39. ICML 2026Left–Right Symmetry Breaking in CLIP-Style Vision-Language Models Trained on Synthetic Spatial-Relation Data
40. ICML 2026Mechanisms of Introspective Awareness
41. ICML 2026Mechanistic Data Attribution: Tracing the Training Origins of Interpretable LLM Units
42. ICML 2026Mechanistic Interpretability as Statistical Estimation: A Variance Analysis
43. ICML 2026Mitigating Hallucinations in Large Vision-Language Models via Causal Route Gating
44. ICML 2026Patterning: The Dual of Interpretability
45. ICML 2026PolySAE: Modeling Feature Interactions in Sparse Autoencoders via Polynomial Decoding
46. ICML 2026Predicting the Emergence of Induction Heads in Language Model Pretraining
47. ICML 2026Protein Circuit Tracing via Cross-layer Transcoders
48. ICML 2026Quantifying LLM Attention-Head Stability: Implications for Circuit Universality
49. ICML 2026Query Circuits: Explaining How Language Models Answer User Prompts
50. ICML 2026Reading Between the Tokens: Improving Preference Predictions through Mechanistic Forecasting

FAQ

How many Mechanistic Interpretability papers were published at NeurIPS, ICLR, ICML and CVPR from 2023 to 2026?

311 in total: 17 in 2023, 49 in 2024, 107 in 2025, 138 in 2026.

How many Mechanistic Interpretability papers are at NeurIPS 2026?

NeurIPS 2026 has 16 Mechanistic Interpretability papers, versus 34 the year before.

How many Mechanistic Interpretability papers are at ICLR 2026?

ICLR 2026 has 46 Mechanistic Interpretability papers, versus 38 the year before.

How many Mechanistic Interpretability papers are at ICML 2026?

ICML 2026 has 74 Mechanistic Interpretability papers, versus 33 the year before.

How many Mechanistic Interpretability papers are at CVPR 2026?

CVPR 2026 has 2 Mechanistic Interpretability papers, versus 2 the year before.

Which conference publishes the most Mechanistic Interpretability papers?

ICML, with 119 papers from 2023 to 2026.