method · first seen 4 Oct 2026
activation patching
Also: causal tracing, causal patching, attention head analysis, causal activation patching, head-level attention analysis, mechanistic interpretability interventions
Interpretability technique that replaces internal activations from one run into another to measure component causal effects.
Mentions per week
PapersCodeDiscussion
Rank 137 on 16 Nov 2026 · 19 mentions in 28 days · 18 independent groups · papers · first seen 6 weeks ago · acceleration +0.80
Earliest sources
- Tracing a Sparse Emotion-Control Circuit in LLM-Based Text-to-SpeechPapers · 4 Oct 2026 · Hongfei Du, Jiacheng Shi, Yanfu Zhang, Ye Gao
- Readable Before Actionable: Causal Tracing of Indirect Prompt InjectionPapers · 4 Oct 2026 · Zhe Yu, Wenpeng Xing, Xingxing Yang, Meng Han
- Separators Make Carry Propagation Learnable:The Geometry of Latent Carry in a Multiplication TransformerPapers · 5 Oct 2026 · Sama Satariyan, Raphael Cousin, G{é}rard Biau
- From Local Evidence to Safety Verdicts: Causal Tracing in Vision-Language ModelsPapers · 5 Oct 2026 · Faezeh Dehghan Tarzjani, Mevan Wijewardena, Alexander Romanus, Sampad Mohanty
- Do Vision-Language-Action Models Understand Instructions? A Mechanistic Interpretability Study on Language GroundingPapers · 7 Oct 2026 · Theodor Wulff, Angelo Cangelosi
- PatchBench: Measuring Collateral Damage in Activation PatchingPapers · 7 Oct 2026 · Alexi Canesse, Mathis Le Bail, Maël Jenny, Clément Elliker, Mahammed El Sharka…
Latest sources
- How VLMs Plan What's NextPapers · 16 Nov 2026 · Rohit Gandikota, Arnab Sen Sharma, Adam Belfki, David Bau
- Seen but Unsaid: Locating and Repairing Cross-Image Comparison in Vision-Language ModelsPapers · 16 Nov 2026 · Zhaoning Zhang
- Programmable Semantic Interventions: Learning Causal Carriers from Behavioral ProgramsPapers · 16 Nov 2026 · Paolo Muratore, Mackenzie Weygandt Mathis
- Before the Word: Name Information in Description-Derived States of Small Language Models, With and Without Naming FailurePapers · 14 Nov 2026 · Fernando Cardenas Piepereit
- Are Causal Neuron Circuits Unique and Transferable?Papers · 10 Nov 2026 · Ömer Tekin Yavuz
- Locating and Repairing Domain Shift in VLM Trajectory PlanningPapers · 7 Nov 2026 · Zhihong Cui, Hengyu Liu, Michael Alexander Riegler, Guandong Xu, Amir Taherkordi…
- Localizing and Editing the Goal Representation of a Vision-Language-Action ModelPapers · 6 Nov 2026 · Sudhanshu Srivastava, Faisal Aqlan
- Language Defaults Before Content: A Causal Layer Map of Multilingual GenerationPapers · 5 Nov 2026 · Shenbin Qian, Yves Scherrer
- When Every Choice Is Correct: A Causal Study of Numerical Preference in Language ModelsPapers · 4 Nov 2026 · Hongnian Huang, Hongliang Liu
- A Circuit for In-Context LearningPapers · 3 Nov 2026 · Andrew Gao, William Gao
- Whose Effect Is It? How Objectives Inflate Learned Subspace InterventionsPapers · 3 Nov 2026 · Toygar Tanyel, Serkan Ayvaz
- Heads That Write, Not Just Point: Image Retrieval Heads in Vision-Language ModelsPapers · 3 Nov 2026 · Junsung Park, Sangwon Yu, Donghun Kang, Yeongtak Oh, Jaewon Jeong, Gyeongtae Yoo…