Atticus Geiger

72 papers A* 12B 2Journal 44Unranked 14

Year	Rank	Type	Title / Venue / Authors
2026	J	`jnl`	From Directions to Regions: Decomposing Activations in Language Models via Local Geometry. CoRR Or Shafran, Shaked Ronen, Omri Fahn, Shauli Ravfogel, Atticus Geiger, Mor Geva
2026	J	`jnl`	Reasoning Theater: Disentangling Model Beliefs from Chain-of-Thought. CoRR Siddharth Boppana, Annabel Ma, Max Loeffler, Raphaël Sarfati, Eric Bigelow, Atticus Geiger, Owen Lewis, Jack Merullo
2026	J	`jnl`	Surgical Activation Steering via Generative Causal Mediation. CoRR Aruna Sankaranarayanan, Amir Zur, Atticus Geiger, Dylan Hadfield-Menell
2026	J	`jnl`	The Shape of Beliefs: Geometry, Dynamics, and Interventions along Representation Manifolds of Language Models' Posteriors. CoRR Raphaël Sarfati, Eric J. Bigelow, Daniel Wurgaft, Jack Merullo, Atticus Geiger, Owen Lewis, Tom McGrath, Ekdeep Singh Lubana
2025	J	`jnl`	Are language models aware of the road not taken? Token-level uncertainty and hidden state dynamics. CoRR Amir Zur, Atticus Geiger, Ekdeep Singh Lubana, Eric J. Bigelow
2025	A*	`conf`	AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders. ICML Zhengxuan Wu, Aryaman Arora, Atticus Geiger, Zheng Wang, Jing Huang, Dan Jurafsky, Christopher D. Manning, Christopher Potts
2025	J	`jnl`	AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders. CoRR Zhengxuan Wu, Aryaman Arora, Atticus Geiger, Zheng Wang, Jing Huang, Dan Jurafsky, Christopher D. Manning, Christopher Potts
2025	J	`jnl`	Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretability. J. Mach. Learn. Res. Atticus Geiger, Duligur Ibeling, Amir Zur, Maheep Chaudhary, Sonakshi Chauhan, Jing Huang, Aryaman Arora, Zhengxuan Wu, Noah D. Goodman, Christopher Potts, Thomas Icard
2025	—	`conf`	Combining Causal Models for More Accurate Abstractions of Neural Networks. CLeaR Theodora-Mara Pîslar, Sara Magliacane, Atticus Geiger
2025	J	`jnl`	Combining Causal Models for More Accurate Abstractions of Neural Networks. CoRR Theodora-Mara Pîslar, Sara Magliacane, Atticus Geiger
2025	J	`jnl`	Decomposing MLP Activations into Interpretable Features via Semi-Nonnegative Matrix Factorization. CoRR Or Shafran, Atticus Geiger, Mor Geva
2025	—	`conf`	Enhancing Automated Interpretability with Output-Centric Feature Descriptions. ACL (1) Yoav Gur-Arieh, Roy Mayan, Chen Agassy, Atticus Geiger, Mor Geva
2025	J	`jnl`	Enhancing Automated Interpretability with Output-Centric Feature Descriptions. CoRR Yoav Gur-Arieh, Roy Mayan, Chen Agassy, Atticus Geiger, Mor Geva
2025	J	`jnl`	How Causal Abstraction Underpins Computational Explanation. CoRR Atticus Geiger, Jacqueline Harding, Thomas Icard
2025	A*	`conf`	How Do Transformers Learn Variable Binding in Symbolic Programs? ICML Yiwei Wu, Atticus Geiger, Raphaël Millière
2025	J	`jnl`	How Do Transformers Learn Variable Binding in Symbolic Programs? CoRR Yiwei Wu, Atticus Geiger, Raphaël Millière
2025	A*	`conf`	HyperDAS: Towards Automating Mechanistic Interpretability with Hypernetworks. ICLR Jiuding Sun, Jing Huang, Sidharth Baskaran, Karel D'Oosterlinck, Christopher Potts, Michael Sklar, Atticus Geiger
2025	J	`jnl`	HyperDAS: Towards Automating Mechanistic Interpretability with Hypernetworks. CoRR Jiuding Sun, Jing Huang, Sidharth Baskaran, Karel D'Oosterlinck, Christopher Potts, Michael Sklar, Atticus Geiger
2025	J	`jnl`	HyperSteer: Activation Steering at Scale with Hypernetworks. CoRR Jiuding Sun, Sidharth Baskaran, Zhengxuan Wu, Michael Sklar, Christopher Potts, Atticus Geiger
2025	J	`jnl`	Language Models use Lookbacks to Track Beliefs. CoRR Nikhil Prakash, Natalie Shapira, Arnab Sen Sharma, Christoph Riedl, Yonatan Belinkov, Tamar Rott Shaham, David Bau, Atticus Geiger
2025	A*	`conf`	MIB: A Mechanistic Interpretability Benchmark. ICML Aaron Mueller, Atticus Geiger, Sarah Wiegreffe, Dana Arad, Iván Arcuschin, Adam Belfki, Yik Siu Chan, Jaden Fried Fiotto-Kaufman, Tal Haklay, Michael Hanna, Jing Huang, Rohan Gupta, Yaniv Nikankin, Hadas Orgad, Nikhil Prakash, Anja Reusch, Aruna Sankaranarayanan, Shun Shao, Alessandro Stolfo, Martin Tutek, Amir Zur, David Bau, Yonatan Belinkov
2025	J	`jnl`	MIB: A Mechanistic Interpretability Benchmark. CoRR Aaron Mueller, Atticus Geiger, Sarah Wiegreffe, Dana Arad, Iván Arcuschin, Adam Belfki, Yik Siu Chan, Jaden Fiotto-Kaufman, Tal Haklay, Michael Hanna, Jing Huang, Rohan Gupta, Yaniv Nikankin, Hadas Orgad, Nikhil Prakash, Anja Reusch, Aruna Sankaranarayanan, Shun Shao, Alessandro Stolfo, Martin Tutek, Amir Zur, David Bau, Yonatan Belinkov
2025	J	`jnl`	Mixing Mechanisms: How Language Models Retrieve Bound Entities In-Context. CoRR Yoav Gur-Arieh, Mor Geva, Atticus Geiger
2025	J	`jnl`	Open Problems in Mechanistic Interpretability. CoRR Lee Sharkey, Bilal Chughtai, Joshua Batson, Jack Lindsey, Jeff Wu, Lucius Bushnaq, Nicholas Goldowsky-Dill, Stefan Heimersheim, Alejandro Ortega, Joseph Isaac Bloom, Stella Biderman, Adrià Garriga-Alonso, Arthur Conmy, Neel Nanda, Jessica Rumbelow, Martin Wattenberg, Nandi Schoots, Joseph Miller, Eric J. Michaud, Stephen Casper, Max Tegmark, William Saunders, David Bau, Eric Todd, Atticus Geiger, Mor Geva, Jesse Hoogland, Daniel Murfet, Tom McGrath
2025	J	`jnl`	Open Problems in Mechanistic Interpretability. Trans. Mach. Learn. Res. Lee Sharkey, Bilal Chughtai, Joshua Batson, Jack Lindsey, Jeffrey Wu, Lucius Bushnaq, Nicholas Goldowsky-Dill, Stefan Heimersheim, Alejandro Ortega, Joseph Isaac Bloom, Stella Biderman, Adrià Garriga-Alonso, Arthur Conmy, Neel Nanda, Jessica Rumbelow, Martin Wattenberg, Nandi Schoots, Joseph Miller, William Saunders, Eric J. Michaud, Stephen Casper, Max Tegmark, David Bau, Eric Todd, Atticus Geiger, Mor Geva, Jesse Hoogland, Daniel Murfet, Tom McGrath
2024	J	`jnl`	A Reply to Makelov et al. (2023)'s "Interpretability Illusion" Arguments. CoRR Zhengxuan Wu, Atticus Geiger, Jing Huang, Aryaman Arora, Thomas Icard, Christopher Potts, Noah D. Goodman
2024	J	`jnl`	Evaluating Open-Source Sparse Autoencoders on Disentangling Factual Knowledge in GPT-2 Small. CoRR Maheep Chaudhary, Atticus Geiger
2024	—	`conf`	Finding Alignments Between Interpretable Causal Variables and Distributed Neural Representations. CLeaR Atticus Geiger, Zhengxuan Wu, Christopher Potts, Thomas Icard, Noah D. Goodman
2024	A*	`conf`	Is This the Subspace You Are Looking for? An Interpretability Illusion for Subspace Activation Patching. ICLR Aleksandar Makelov, Georg Lange, Atticus Geiger, Neel Nanda
2024	—	`conf`	RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations. ACL (1) Jing Huang, Zhengxuan Wu, Christopher Potts, Mor Geva, Atticus Geiger
2024	J	`jnl`	RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations. CoRR Jing Huang, Zhengxuan Wu, Christopher Potts, Mor Geva, Atticus Geiger
2024	A*	`conf`	ReFT: Representation Finetuning for Language Models. NeurIPS Zhengxuan Wu, Aryaman Arora, Zheng Wang, Atticus Geiger, Dan Jurafsky, Christopher D. Manning, Christopher Potts
2024	J	`jnl`	ReFT: Representation Finetuning for Language Models. CoRR Zhengxuan Wu, Aryaman Arora, Zheng Wang, Atticus Geiger, Dan Jurafsky, Christopher D. Manning, Christopher Potts
2024	J	`jnl`	Recurrent Neural Networks Learn to Store and Generate Sequences using Non-Linear Representations. CoRR Róbert Csordás, Christopher Potts, Christopher D. Manning, Atticus Geiger
2024	A*	`conf`	Updating CLIP to Prefer Descriptions Over Captions. EMNLP Amir Zur, Elisa Kreiss, Karel D'Oosterlinck, Christopher Potts, Atticus Geiger
2024	J	`jnl`	Updating CLIP to Prefer Descriptions Over Captions. CoRR Amir Zur, Elisa Kreiss, Karel D'Oosterlinck, Christopher Potts, Atticus Geiger
2024	—	`conf`	pyvene: A Library for Understanding and Improving PyTorch Models via Interventions. NAACL (Demonstrations) Zhengxuan Wu, Atticus Geiger, Aryaman Arora, Jing Huang, Zheng Wang, Noah D. Goodman, Christopher D. Manning, Christopher Potts
2024	J	`jnl`	pyvene: A Library for Understanding and Improving PyTorch Models via Interventions. CoRR Zhengxuan Wu, Atticus Geiger, Aryaman Arora, Jing Huang, Zheng Wang, Noah D. Goodman, Christopher D. Manning, Christopher Potts
2023	B	`conf`	A Semantics for Causing, Enabling, and Preventing Verbs Using Structural Causal Models. CogSci Angela Cao, Atticus Geiger, Elisa Kreiss, Thomas Icard, Tobias Gerstenberg
2023	J	`jnl`	Causal Abstraction for Faithful Model Interpretation. CoRR Atticus Geiger, Christopher Potts, Thomas Icard
2023	—	`conf`	Causal Abstraction with Soft Interventions. CLeaR Riccardo Massidda, Atticus Geiger, Thomas Icard, Davide Bacciu
2023	A*	`conf`	Causal Proxy Models for Concept-based Model Explanations. ICML Zhengxuan Wu, Karel D'Oosterlinck, Atticus Geiger, Amir Zur, Christopher Potts
2023	J	`jnl`	Finding Alignments Between Interpretable Causal Variables and Distributed Neural Representations. CoRR Atticus Geiger, Zhengxuan Wu, Christopher Potts, Thomas Icard, Noah D. Goodman
2023	A*	`conf`	Interpretability at Scale: Identifying Causal Mechanisms in Alpaca. NeurIPS Zhengxuan Wu, Atticus Geiger, Thomas Icard, Christopher Potts, Noah D. Goodman
2023	J	`jnl`	Interpretability at Scale: Identifying Causal Mechanisms in Alpaca. CoRR Zhengxuan Wu, Atticus Geiger, Christopher Potts, Noah D. Goodman
2023	J	`jnl`	Linear Representations of Sentiment in Large Language Models. CoRR Curt Tigges, Oskar John Hollinsworth, Atticus Geiger, Neel Nanda
2023	—	`conf`	Rigorously Assessing Natural Language Explanations of Neurons. BlackboxNLP@EMNLP Jing Huang, Atticus Geiger, Karel D'Oosterlinck, Zhengxuan Wu, Christopher Potts
2023	J	`jnl`	Rigorously Assessing Natural Language Explanations of Neurons. CoRR Jing Huang, Atticus Geiger, Karel D'Oosterlinck, Zhengxuan Wu, Christopher Potts
2023	—	`conf`	ScoNe: Benchmarking Negation Reasoning in Language Models With Fine-Tuning and In-Context Learning. ACL (2) Jingyuan Selena She, Christopher Potts, Samuel R. Bowman, Atticus Geiger
2023	J	`jnl`	ScoNe: Benchmarking Negation Reasoning in Language Models With Fine-Tuning and In-Context Learning. CoRR Jingyuan Selena She, Christopher Potts, Samuel R. Bowman, Atticus Geiger
2022	A*	`conf`	CEBaB: Estimating the Causal Effects of Real-World Concepts on NLP Model Behavior. NeurIPS Eldar David Abraham, Karel D'Oosterlinck, Amir Feder, Yair Ori Gat, Atticus Geiger, Christopher Potts, Roi Reichart, Zhengxuan Wu
2022	J	`jnl`	CEBaB: Estimating the Causal Effects of Real-World Concepts on NLP Model Behavior. CoRR Eldar David Abraham, Karel D'Oosterlinck, Amir Feder, Yair Ori Gat, Atticus Geiger, Christopher Potts, Roi Reichart, Zhengxuan Wu
2022	J	`jnl`	Causal Abstraction with Soft Interventions. CoRR Riccardo Massidda, Atticus Geiger, Thomas Icard, Davide Bacciu
2022	—	`conf`	Causal Distillation for Language Models. NAACL-HLT Zhengxuan Wu, Atticus Geiger, Joshua Rozner, Elisa Kreiss, Hanson Lu, Thomas Icard, Christopher Potts, Noah D. Goodman
2022	J	`jnl`	Causal Proxy Models for Concept-Based Model Explanations. CoRR Zhengxuan Wu, Karel D'Oosterlinck, Atticus Geiger, Amir Zur, Christopher Potts
2022	A*	`conf`	Inducing Causal Structure for Interpretable Neural Networks. ICML Atticus Geiger, Zhengxuan Wu, Hanson Lu, Josh Rozner, Elisa Kreiss, Thomas Icard, Noah D. Goodman, Christopher Potts
2021	A*	`conf`	Causal Abstractions of Neural Networks. NeurIPS Atticus Geiger, Hanson Lu, Thomas Icard, Christopher Potts
2021	J	`jnl`	Causal Abstractions of Neural Networks. CoRR Atticus Geiger, Hanson Lu, Thomas Icard, Christopher Potts
2021	J	`jnl`	Causal Distillation for Language Models. CoRR Zhengxuan Wu, Atticus Geiger, Josh Rozner, Elisa Kreiss, Hanson Lu, Thomas Icard, Christopher Potts, Noah D. Goodman
2021	—	`conf`	DynaSent: A Dynamic Benchmark for Sentiment Analysis. ACL/IJCNLP (1) Christopher Potts, Zhengxuan Wu, Atticus Geiger, Douwe Kiela
2021	—	`conf`	Dynabench: Rethinking Benchmarking in NLP. NAACL-HLT Douwe Kiela, Max Bartolo, Yixin Nie, Divyansh Kaushik, Atticus Geiger, Zhengxuan Wu, Bertie Vidgen, Grusha Prasad, Amanpreet Singh, Pratik Ringshia, Zhiyi Ma, Tristan Thrush, Sebastian Riedel, Zeerak Waseem, Pontus Stenetorp, Robin Jia, Mohit Bansal, Christopher Potts, Adina Williams
2021	J	`jnl`	Dynabench: Rethinking Benchmarking in NLP. CoRR Douwe Kiela, Max Bartolo, Yixin Nie, Divyansh Kaushik, Atticus Geiger, Zhengxuan Wu, Bertie Vidgen, Grusha Prasad, Amanpreet Singh, Pratik Ringshia, Zhiyi Ma, Tristan Thrush, Sebastian Riedel, Zeerak Waseem, Pontus Stenetorp, Robin Jia, Mohit Bansal, Christopher Potts, Adina Williams
2021	J	`jnl`	Inducing Causal Structure for Interpretable Neural Networks. CoRR Atticus Geiger, Zhengxuan Wu, Hanson Lu, Josh Rozner, Elisa Kreiss, Thomas Icard, Noah D. Goodman, Christopher Potts
2020	J	`jnl`	DynaSent: A Dynamic Benchmark for Sentiment Analysis. CoRR Christopher Potts, Zhengxuan Wu, Atticus Geiger, Douwe Kiela
2020	J	`jnl`	Modular Representation Underlies Systematic Generalization in Neural Natural Language Inference Models. CoRR Atticus Geiger, Kyle Richardson, Christopher Potts
2020	—	`conf`	Neural Natural Language Inference Models Partially Embed Theories of Lexical Entailment and Negation. BlackboxNLP@EMNLP Atticus Geiger, Kyle Richardson, Christopher Potts
2020	B	`conf`	Relational reasoning and generalization using non-symbolic neural networks. CogSci Atticus Geiger, Alexandra Carstensen, Michael C. Frank, Christopher Potts
2020	J	`jnl`	Relational reasoning and generalization using non-symbolic neural networks. CoRR Atticus Geiger, Alexandra Carstensen, Michael C. Frank, Christopher Potts
2019	—	`conf`	Posing Fair Generalization Tasks for Natural Language Inference. EMNLP/IJCNLP (1) Atticus Geiger, Ignacio Cases, Lauri Karttunen, Christopher Potts
2019	J	`jnl`	Posing Fair Generalization Tasks for Natural Language Inference. CoRR Atticus Geiger, Ignacio Cases, Lauri Karttunen, Christopher Potts
2019	—	`conf`	Recursive Routing Networks: Learning to Compose Modules for Language Understanding. NAACL-HLT (1) Ignacio Cases, Clemens Rosenbaum, Matthew Riemer, Atticus Geiger, Tim Klinger, Alex Tamkin, Olivia Li, Sandhini Agarwal, Joshua D. Greene, Dan Jurafsky, Christopher Potts, Lauri Karttunen
2018	J	`jnl`	Stress-Testing Neural Models of Natural Language Inference with Multiply-Quantified Sentences. CoRR Atticus Geiger, Ignacio Cases, Lauri Karttunen, Christopher Potts