Methods › Natural Language Processing › Inference Extrapolation › Patching
Activation Patching
Patching
Introduced by Asma Ghandeharioun et al. in Patchscopes: A Unifying Framework for Inspecting Hidden Representations of Language Models
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
Activation patching studies the model's computation by altering its latent representations, the token embeddings in transformer-based language models, during the inference process
In Syntology name not resolved (unknown_name). The lookup matched no method, paper or venue by this name; that asserts nothing about whether code exists.
Near matches the lookup lists, each a different name: back-patching; self-patching; local patching; patching method; patching strategy; Patching algorithm; direction patching; attribution patching.
Syntology's have check for this name (free, it answers coverage only), read 2026-10-05; ask it again live. A level is the served verification level of that one answer; nothing here is summed across methods.
Papers archive 2025-07-28
archive papers tagged: 102 · with a code link: 43 · where Syntology ran a sample: 19 (11 with a run with no instrument failure, 8 where every run was a failure of Syntology's instrument) Syntology
Show: all tagged papersonly where code ran (19 of 102 tagged: 11 with a run with no instrument failure, 8 where every run was a failure of Syntology's instrument)
Page 1 of 2: papers 1 to 100 of 102, newest first by the archive's date (ties by slug), in archive order.
Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code, as “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. It is per sample and not a correctness claim. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from.
-
Data Augmentation in Time Series Forecasting through Inverted Framework 15 Jul 2025 · 0 repositories · arXiv:2507.11439
-
Adversarial Activation Patching: A Framework for Detecting and Mitigating Emergent Deception in Safety-Aligned Transformers 12 Jul 2025 · 0 repositories · arXiv:2507.09406
-
Unpatchable Vulnerabilities in Windows 10/11: Security Report 2025 10 Jul 2025 · 0 repositories
-
Q-resafe: Assessing Safety Risks and Quantization-aware Safety Patching for Quantized Large Language Models 25 Jun 2025 · 1 repository · arXiv:2506.20251Syntology official (archive's flag): 2 ran · 2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified (of 2 harvested samples) · 2 pointer-only (licence)
-
SEC-bench: Automated Benchmarking of LLM Agents on Real-World Software Security Tasks 13 Jun 2025 · 1 repository · arXiv:2506.11791
-
Time Series Representations for Classification Lie Hidden in Pretrained Vision Transformers 10 Jun 2025 · 0 repositories · arXiv:2506.08641
-
Chasing Moving Targets with Online Self-Play Reinforcement Learning for Safer Language Models 9 Jun 2025 · 1 repository · arXiv:2506.07468Syntology official (archive's flag): 5 ran · 6 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 5 where Syntology's instrument failed) · 11 unverified (of 17 harvested samples) · 1 pointer-only (licence)
-
A Multi-Dataset Evaluation of Models for Automated Vulnerability Repair 5 Jun 2025 · 0 repositories · arXiv:2506.04987
-
Can LLMs Reason Abstractly Over Math Word Problems Without CoT? Disentangling Abstract Formulation From Arithmetic Computation 29 May 2025 · 0 repositories · arXiv:2505.23701
-
Foundation Model for Wireless Technology Recognition Using IQ Timeseries 26 May 2025 · 0 repositories · arXiv:2505.19390
-
From What to How: Attributing CLIP's Latent Components Reveals Unexpected Semantic Reliance 26 May 2025 · 1 repository · arXiv:2505.20229
-
Co-PatcheR: Collaborative Software Patching with Component(s)-specific Small Reasoning Models 25 May 2025 · 0 repositories · arXiv:2505.18955
-
Know the Ropes: A Heuristic Strategy for LLM-based Multi-Agent System Design 22 May 2025 · 0 repositories · arXiv:2505.16979
-
Internal Chain-of-Thought: Empirical Evidence for Layer-wise Subtask Scheduling in LLMs 20 May 2025 · 1 repository · arXiv:2505.14530
-
SPIRIT: Patching Speech Language Models against Jailbreak Attacks 18 May 2025 · 0 repositories · arXiv:2505.13541
-
SPAT: Sensitivity-based Multihead-attention Pruning on Time Series Forecasting Models 13 May 2025 · 0 repositories · arXiv:2505.08768
-
Unpacking Robustness in Inflectional Languages: Adversarial Evaluation and Mechanistic Insights 8 May 2025 · 0 repositories · arXiv:2505.07856
-
Interpreting Multilingual and Document-Length Sensitive Relevance Computations in Neural Retrieval Models through Axiomatic Causal Interventions 4 May 2025 · 1 repository · arXiv:2505.02154
-
The Illusion of Role Separation: Hidden Shortcuts in LLM Role Learning (and How to Fix Them) 1 May 2025 · 0 repositories · arXiv:2505.00626
-
HSE: A plug-and-play module for unified fault diagnosis foundation models 26 Apr 2025 · 1 repository
-
MIB: A Mechanistic Interpretability Benchmark 17 Apr 2025 · 1 repository · arXiv:2504.13151Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified (of 3 harvested samples)
-
How do Large Language Models Understand Relevance? A Mechanistic Interpretability Perspective 10 Apr 2025 · 1 repository · arXiv:2504.07898
-
Localized Definitions and Distributed Reasoning: A Proof-of-Concept Mechanistic Interpretability Study via Activation Patching 3 Apr 2025 · 1 repository · arXiv:2504.02976
-
Reverse-Engineering the Retrieval Process in GenIR Models 25 Mar 2025 · 1 repository · arXiv:2503.19715
-
Sentinel: Multi-Patch Transformer with Temporal and Channel Attention for Time Series Forecasting 22 Mar 2025 · 0 repositories · arXiv:2503.17658
-
A Semantic-based Optimization Approach for Repairing LLMs: Case Study on Code Generation 17 Mar 2025 · 0 repositories · arXiv:2503.12899
-
TinySQL: A Progressive Text-to-SQL Dataset for Mechanistic Interpretability Research 17 Mar 2025 · 0 repositories · arXiv:2503.12730
-
Rethinking Lanes and Points in Complex Scenarios for Monocular 3D Lane Detection 8 Mar 2025 · 0 repositories · arXiv:2503.06237
-
TimeFound: A Foundation Model for Time Series Forecasting 6 Mar 2025 · 0 repositories · arXiv:2503.04118
-
Superscopes: Amplifying Internal Feature Representations for Language Model Interpretation 3 Mar 2025 · 1 repository · arXiv:2503.02078
-
To Patch or Not to Patch: Motivations, Challenges, and Implications for Cybersecurity 24 Feb 2025 · 0 repositories · arXiv:2502.17703
-
Exploring Translation Mechanism of Large Language Models 17 Feb 2025 · 0 repositories · arXiv:2502.11806
-
RomanLens: Latent Romanization and its role in Multilinguality in LLMs 11 Feb 2025 · 0 repositories · arXiv:2502.07424
-
EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification 7 Feb 2025 · 0 repositories · arXiv:2502.06852
-
Decoding Human Attentive States from Spatial-temporal EEG Patches Using Transformers 6 Feb 2025 · 1 repository · arXiv:2502.03736
-
PatchPilot: A Cost-Efficient Software Engineering Agent with Early Attempts on Formal Verification 4 Feb 2025 · 1 repository · arXiv:2502.02747
-
Rethinking Vision Transformer for Object Centric Foundation Models 4 Feb 2025 · 1 repository · arXiv:2502.02763
-
Multi-Grained Patch Training for Efficient LLM-based Recommendation 25 Jan 2025 · 1 repository · arXiv:2501.15087Syntology official (archive's flag): 6 ran · 6 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified (of 6 harvested samples)
-
Surface Vision Mamba: Leveraging Bidirectional State Space Model for Efficient Spherical Manifold Representation 24 Jan 2025 · 0 repositories · arXiv:2501.14679
-
Using Pre-trained LLMs for Multivariate Time Series Forecasting 10 Jan 2025 · 0 repositories · arXiv:2501.06386
-
xPatch: Dual-Stream Time Series Forecasting with Exponential Seasonal-Trend Decomposition 23 Dec 2024 · 1 repository · arXiv:2412.17323
-
WPMixer: Efficient Multi-Resolution Mixing for Long-Term Time Series Forecasting 22 Dec 2024 · 1 repository · arXiv:2412.17176Syntology official (archive's flag): 5 ran · 5 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 2 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified (of 5 harvested samples) · 3 pointer-only (licence)
-
Effective and Efficient Representation Learning for Flight Trajectories 21 Dec 2024 · 1 repository · arXiv:2412.16581
-
Understanding Individual Agent Importance in Multi-Agent System via Counterfactual Reasoning 20 Dec 2024 · 0 repositories · arXiv:2412.15619
-
Benchmarking and Understanding Compositional Relational Reasoning of LLMs 17 Dec 2024 · 1 repository · arXiv:2412.12841Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 4 unverified (of 5 harvested samples) · 5 pointer-only (licence)
-
Client-Side Patching against Backdoor Attacks in Federated Learning 13 Dec 2024 · 0 repositories · arXiv:2412.10605
-
Efficient Large-Scale Traffic Forecasting with Transformers: A Spatial Data Management Perspective 13 Dec 2024 · 3 repositories · arXiv:2412.09972Syntology official (archive's flag): 3 ran · 5 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 1 violated, 2 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified (of 6 harvested samples)
-
A New Perspective on Time Series Anomaly Detection: Faster Patch-based Broad Learning System 7 Dec 2024 · 0 repositories · arXiv:2412.05498
-
Transforming Engineering Diagrams: A Novel Approach for P&ID Digitization using Transformers 21 Nov 2024 · 0 repositories · arXiv:2411.13929
-
UniFlow: A Foundation Model for Unified Urban Spatio-Temporal Flow Prediction 20 Nov 2024 · 1 repository · arXiv:2411.12972
-
SAG-ViT: A Scale-Aware, High-Fidelity Patching Approach with Graph Attention for Vision Transformers 14 Nov 2024 · 1 repository · arXiv:2411.09420
-
Separating Tongue from Thought: Activation Patching Reveals Language-Agnostic Concept Representations in Transformers 13 Nov 2024 · 1 repository · arXiv:2411.08745Syntology official (archive's flag): 6 ran · 6 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 5 where Syntology's instrument failed) · 4 unverified (of 10 harvested samples) · 10 pointer-only (licence)
-
Beyond Toxic Neurons: A Mechanistic Analysis of DPO for Toxicity Reduction 10 Nov 2024 · 1 repository · arXiv:2411.06424Syntology official (archive's flag): 13 ran · 13 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 1 honoured, 0 violated, 10 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified (of 15 harvested samples)
-
Brain Tumour Removing and Missing Modality Generation using 3D WDM 7 Nov 2024 · 1 repository · arXiv:2411.04630
-
FlexTSF: A Universal Forecasting Model for Time Series with Variable Regularities 30 Oct 2024 · 1 repository · arXiv:2410.23160
-
Exploring Tokenization Methods for Multitrack Sheet Music Generation 23 Oct 2024 · 0 repositories · arXiv:2410.17584
-
Graph Neural Patching for Cold-Start Recommendations 18 Oct 2024 · 0 repositories · arXiv:2410.14241
-
Dissecting Fine-Tuning Unlearning in Large Language Models 9 Oct 2024 · 1 repository · arXiv:2410.06606
-
Enhanced Transformer architecture for in-context learning of dynamical systems 4 Oct 2024 · 2 repositories · arXiv:2410.03291
-
VulnLLMEval: A Framework for Evaluating Large Language Models in Software Vulnerability Detection and Patching 16 Sep 2024 · 0 repositories · arXiv:2409.10756
-
Extracting Paragraphs from LLM Token Activations 10 Sep 2024 · 0 repositories · arXiv:2409.06328
-
Representational Analysis of Binding in Language Models 9 Sep 2024 · 0 repositories · arXiv:2409.05448
-
Enhancing Automated Program Repair with Solution Design 22 Aug 2024 · 0 repositories · arXiv:2408.12056
-
Unlocking the Power of LSTM for Long Term Time Series Forecasting 19 Aug 2024 · 0 repositories · arXiv:2408.10006
-
PatUntrack: Automated Generating Patch Examples for Issue Reports without Tracked Insecure Code 16 Aug 2024 · 0 repositories · arXiv:2408.08619
-
MambaST: A Plug-and-Play Cross-Spectral Spatial-Temporal Fuser for Efficient Pedestrian Detection 2 Aug 2024 · 1 repository · arXiv:2408.01037
-
A federated large language model for long-term time series forecasting 30 Jul 2024 · 0 repositories · arXiv:2407.20503
-
Answer, Assemble, Ace: Understanding How Transformers Answer Multiple Choice Questions 21 Jul 2024 · 0 repositories · arXiv:2407.15018Syntology 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified (of 1 harvested sample)
-
Investigating the Indirect Object Identification circuit in Mamba 19 Jul 2024 · 1 repository · arXiv:2407.14008
-
Mechanistic interpretability of large language models with applications to the financial services industry 15 Jul 2024 · 0 repositories · arXiv:2407.11215
-
MoreFixes: A Large-Scale Dataset of CVE Fix Commits Mined through Enhanced Repository Discovery 10 Jul 2024 · 1 repository
-
Functional Faithfulness in the Wild: Circuit Discovery with Differentiable Computation Graph Pruning 4 Jul 2024 · 0 repositories · arXiv:2407.03779
-
Multiple-Resolution Tokenization for Time Series Forecasting with an Application to Pricing 3 Jul 2024 · 1 repository · arXiv:2407.03185
-
SwiftDiffusion: Efficient Diffusion Model Serving with Add-on Modules 2 Jul 2024 · 0 repositories · arXiv:2407.02031
-
Efficient Automated Circuit Discovery in Transformers using Contextual Decomposition 1 Jul 2024 · 0 repositories · arXiv:2407.00886
-
On the Role of Long-tail Knowledge in Retrieval Augmented Large Language Models 24 Jun 2024 · 0 repositories · arXiv:2406.16367
-
Distributed Rule Vectors is A Key Mechanism in Large Language Models' In-Context Learning 23 Jun 2024 · 0 repositories · arXiv:2406.16007
-
Are Language Models Actually Useful for Time Series Forecasting? 22 Jun 2024 · 3 repositories · arXiv:2406.16964Syntology official (archive's flag): 2 ran · 5 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 3 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified (of 5 harvested samples) · 2 pointer-only (licence)
-
Modular Pluralism: Pluralistic Alignment via Multi-LLM Collaboration 22 Jun 2024 · 1 repository · arXiv:2406.15951Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified (of 2 harvested samples) · 2 pointer-only (licence)
-
Finding Safety Neurons in Large Language Models 20 Jun 2024 · 0 repositories · arXiv:2406.14144Syntology 14 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 8 where Syntology's instrument failed) · 0 unverified (of 14 harvested samples) · 1 pointer-only (licence)
-
Deep Learning methodology for the identification of wood species using high-resolution macroscopic images 17 Jun 2024 · 1 repository · arXiv:2406.11772
-
Continuum Attention for Neural Operators 10 Jun 2024 · 1 repository · arXiv:2406.06486Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified (of 4 harvested samples) · 4 pointer-only (licence)
-
AROMA: Preserving Spatial Structure for Latent PDE Modeling with Local Neural Fields 4 Jun 2024 · 1 repository · arXiv:2406.02176Syntology official (archive's flag): 7 ran · 7 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 2 honoured, 2 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 3 unverified (of 10 harvested samples) · 2 pointer-only (licence)
-
Diffusion Model Patching via Mixture-of-Prompts 28 May 2024 · 0 repositories · arXiv:2405.17825
-
No Two Devils Alike: Unveiling Distinct Mechanisms of Fine-tuning Attacks 25 May 2024 · 0 repositories · arXiv:2405.16229
-
Medformer: A Multi-Granularity Patching Transformer for Medical Time-Series Classification 24 May 2024 · 2 repositories · arXiv:2405.19363Syntology official (archive's flag): 8 ran · 8 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 7 where Syntology's instrument failed) · 0 unverified (of 8 harvested samples) · 2 pointer-only (licence)
-
SPLICE -- Streamlining Digital Pathology Image Processing 26 Apr 2024 · 0 repositories · arXiv:2404.17704
-
How to use and interpret activation patching 23 Apr 2024 · 0 repositories · arXiv:2404.15255
-
Explainable Survival Analysis with Uncertainty using Convolution-Involved Vision Transformer 18 Apr 2024 · 0 repositories
-
Adaptive Patching for High-resolution Image Segmentation with Transformers 15 Apr 2024 · 0 repositories · arXiv:2404.09707
-
Unveiling LLMs: The Evolution of Latent Representations in a Dynamic Knowledge Graph 4 Apr 2024 · 1 repository · arXiv:2404.03623Syntology official (archive's flag): 5 ran · 5 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified (of 5 harvested samples) · 5 pointer-only (licence)
-
Harnessing The Power of Attention For Patch-Based Biomedical Image Classification 1 Apr 2024 · 0 repositories · arXiv:2404.00949
-
Have Faith in Faithfulness: Going Beyond Circuit Overlap When Finding Model Mechanisms 26 Mar 2024 · 3 repositories · arXiv:2403.17806Syntology official (archive's flag): 2 ran · 2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified (of 2 harvested samples) · 2 pointer-only (licence)
-
Monotonic Representation of Numeric Properties in Language Models 15 Mar 2024 · 1 repository · arXiv:2403.10381Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified (of 4 harvested samples)
-
AtP*: An efficient and scalable method for localizing LLM behaviour to components 1 Mar 2024 · 0 repositories · arXiv:2403.00745
-
Cutting Off the Head Ends the Conflict: A Mechanism for Interpreting and Mitigating Knowledge Conflicts in Language Models 28 Feb 2024 · 0 repositories · arXiv:2402.18154
-
Fine-Tuning Enhances Existing Mechanisms: A Case Study on Entity Tracking 22 Feb 2024 · 0 repositories · arXiv:2402.14811
-
Mudjacking: Patching Backdoor Vulnerabilities in Foundation Models 22 Feb 2024 · 0 repositories · arXiv:2402.14977
-
Attention as Robust Representation for Time Series Forecasting 8 Feb 2024 · 0 repositories · arXiv:2402.05370
-
Sparse-VQ Transformer: An FFN-Free Framework with Vector Quantization for Enhanced Time Series Forecasting 8 Feb 2024 · 0 repositories · arXiv:2402.05830
Tasks archive 2025-07-28
20 shown of 118 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.
Usage over time archive 2025-07-28
Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).
Categories archive 2025-07-28
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections