Here are the accepted posters for the Workshop on Actionable Interpretability 2025:
Session 1 (10:40 - 11:40)
1 - Benchmarking Deception Probes via Black-to-White Performance Boosts
Avi Parrack, Carlo Leonardo Attubato, Stefan Heimersheim
Poster Session 1
8 - Avoiding Leakage Poisoning: Concept Interventions Under Distribution Shifts
Mateo Espinosa Zarlenga, Gabriele Dominici, Pietro Barbiero, Zohreh Shams, Mateja Jamnik
Poster Session 1
20 - Taming Knowledge Conflicts in Language Models
Gaotang Li, Yuzhong Chen, Hanghang Tong
Poster Session 1
22 - Looking Beyond The Top-1: Transformers Determine Top Tokens In Order
Daria Lioubashevski, Tomer M. Schlank, Gabriel Stanovsky, Ariel Goldstein
Poster Session 1
31 - SAEs Can Improve Unlearning: Dynamic Sparse Autoencoder Guardrails for Precision Unlearning in LLMs
Aashiq Muhamed, Jacopo Bonato, Mona T. Diab, Virginia Smith
Poster Session 1
35 - Supernova Event Dataset: Interpreting Large Language Models' Personality through Critical Event Analysis
Pranav Agarwal, Ioana Ciucă
Poster Session 1
46 - Insights into a radiology-specialised multimodal large language model with sparse autoencoders
Kenza Bouzid, Shruthi Bannur, Felix Meissen, Daniel C. Castro, Anton Schwaighofer, Javier Alvarez-Valle, Stephanie Hyland
Poster Session 1
49 - Needle in a Patched Haystack: Evaluating Saliency Maps for Vision LLMs.
Bastien Zimmermann, Matthieu Boussard
Poster Session 1
51 - Internal Causal Mechanisms Robustly Predict Language Model Out-of-Distribution Behaviors
Jing Huang, Junyi Tao, Thomas Icard, Diyi Yang, Christopher Potts
Poster Session 1
53 - Single Feature Tips the Balance: Reducing Language Model Over-Refusal with Sparse Representations
Ailin Deng, Shaoliang Nie, Lijuan Liu, Xianjun Yang, Ujjwal Karn, Dat Huynh, Fulton Wang, Ying Xu, Madian Khabsa, Saghar Hosseini
Poster Session 1
55 - Steering Language Model Refusal with Sparse Autoencoders
Kyle O'Brien, David Majercak, Xavier Fernandes, Richard G. Edgar, Blake Bullwinkel, Jingya Chen, Harsha Nori, Dean Carignan, Eric Horvitz, Forough Poursabzi-Sangdeh
Poster Session 1
56 - Koopman Autoencoders Learn Neural Representation Dynamics
Nishant Suresh Aswani, Saif Jabari
Poster Session 1
75 - Scalpel vs. Hammer: GRPO Amplifies Existing Capabilities, SFT Replaces Them
Neel Rajani, Aryo Pradipta Gema, Seraphina Goldfarb-Tarrant, Ivan Titov
Poster Session 1
78 - Pruning the Paradox: How CLIP’s Most Informative Heads Enhance Performance While Amplifying Bias
Avinash Madasu, Vasudev Lal, Phillip Howard
Poster Session 1
85 - Actionable Interpretability via Causal Hypergraphs: Unravelling Batch Size Effects in Deep Learning
Zhongtian Sun, Anoushka Harit, Pietro Lio
Poster Session 1
86 - Reasoning-Finetuning Repurposes Latent Mechanisms in Base Models
Jake Ward, Chuqiao Lin
Poster Session 1
87 - MIB: A Mechanistic Interpretability Benchmark
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
Poster Session 1
88 - When Compression Breaks Safety, Interpretability Reveals The Fix
Vishnu Kabir Chhabra, Mohammad Mahdi Khalili
Poster Session 1
91 - Challenges in Understanding Modality Conflict in Vision-Language Models
Trang Nguyen, Jackson Sam Michaels, Madalina Fiterau, David Jensen
Poster Session 1
130 - Interpretable Diffusion Models with B-cos Networks
Nicola Andri Bernold, Moritz Vandenhirtz, Alice Bizeul, Julia E Vogt
Poster Session 1
135 - Detecting High-Stakes Interactions with Activation Probes
Alex McKenzie, Phil Blandfort, Urja Pawar, William Bankes, David Krueger, Ekdeep Singh Lubana, Dmitrii Krasheninnikov
Poster Session 1
138 - Probabilistic Soundness Guarantees in LLM Reasoning Chains
Weiqiu You, Anton Xue, Shreya Havaldar, Delip Rao, Helen Jin, Chris Callison-Burch, Eric Wong
Poster Session 1
139 - Probing for Arithmetic Errors in Language Models
Yucheng Sun, Alessandro Stolfo, Mrinmaya Sachan
Poster Session 1
150 - Classifier Reconstruction Through Counterfactual-Aware Wasserstein Prototypes
Xuan Zhao, Zhuo Cao, Arya Bangun, Hanno Scharr, Ira Assent
Poster Session 1
156 - MSE-Break: Steering Internal Representations to Bypass Refusals in Large Language Models
Ashwin Saraswatula, Pranav Balabhadra, Pranav Dhinakar
Poster Session 1
159 - Simplicity is not as simple as you think
Aman Sinha, Timothee Mickus, Marianne Clausel, Mathieu Constant, Xavier Coubez
Poster Session 1
160 - Spectral Scaling Laws in Language Models: How Effectively Do Feed-Forward Networks Use Their Latent Space?
Nandan Kumar Jha, Brandon Reagen
Poster Session 1
Session 2 (13:00 - 14:00)
6 - Unifying Image Counterfactuals and Feature Attributions with Latent-Space Adversarial Attacks
Jeremy Goldwasser, Giles Hooker
Poster Session 2
14 - Sum-of-Parts: Self-Attributing Neural Networks with End-to-End Learning of Feature Groups
Weiqiu You, Helen Qu, Marco Gatti, Bhuvnesh Jain, Eric Wong
Poster Session 2
16 - Learning interpretable positional encodings in transformers depends on initialization
Takuya Ito, Luca Cocchi, Tim Klinger, Parikshit Ram, Murray Campbell, Luke J. Hearne
Poster Session 2
21 - Interpreting the repeated token phenomenon in LLMs
Itay Yona, Jamie Hayes, Ilia Shumailov, Federico Barbero, Yossi Gandelsman
Poster Session 2
32 - The Geometry of Forgetting: Analyzing Machine Unlearning through Local Learning Coefficients
Aashiq Muhamed, Virginia Smith
Poster Session 2
37 - Probing and Steering Evaluation Awareness of Language Models
Jord Nguyen, Hoang Huu Khiem, Carlo Leonardo Attubato, Felix Hofstätter
Poster Session 2
52 - Steering off Course: Reliability Challenges in Steering Language Models
Patrick Queiroz Da Silva, Hari Sethuraman, Dheeraj Rajagopal, Hannaneh Hajishirzi, Sachin Kumar
Poster Session 2
62 - Beyond Multiple Choice: Evaluating Steering Vectors for Adaptive Free-Form Summarization
Joschka Braun, Carsten Eickhoff, Seyed Ali Bahrainian
Poster Session 2
66 - Do LLMs Lie About What They Use? Benchmark for Metacognition Truthfulness in Large Language Models
Nhi Nguyen, Shauli Ravfogel, Rajesh Ranganath
Poster Session 2
66 - Do LLMs Lie About What They Use? Benchmark for Metacognitive Truthfulness in Large Language Models
Nhi Nguyen, Shauli Ravfogel, Rajesh Ranganath
Poster Session 2
68 - Activation Steering in Generative Settings via Contrastive Causal Mediation Analysis
Aruna Sankaranarayanan, Amir Zur, Atticus Geiger, Dylan Hadfield-Menell
Poster Session 2
76 - Towards eliciting latent knowledge from LLMs with mechanistic interpretability
Bartosz Cywiński, Emil Ryd, Senthooran Rajamanoharan, Neel Nanda
Poster Session 2
80 - Beyond the ATE: Interpretable Modelling of Treatment Effects over Dose and Time
Julianna Piskorz, Krzysztof Kacprzyk, Harry Amad, Mihaela van der Schaar
Poster Session 2
81 - Neural Concept Verifier: Scaling Prover-Verifier Games via Concept Encodings
Berkant Turan, Suhrab Asadulla, David Steinmann, Wolfgang Stammer, Sebastian Pokutta
Poster Session 2
84 - Understanding Synthetic Context Extension via Retrieval Heads
Xinyu Zhao, Fangcong Yin, Greg Durrett
Poster Session 2
92 - SAeUron: Interpretable Concept Unlearning in Diffusion Models with Sparse Autoencoders
Bartosz Cywiński, Kamil Deja
Poster Session 2
94 - Automated Detection of Visual Attribute Reliance with a Self-Reflective Agent
Christy Li, Josep Lopez Camuñas, Jake Thomas Touchet, Jacob Andreas, Agata Lapedriza, Antonio Torralba, Tamar Rott Shaham
Poster Session 2
99 - Evaluating Neuron Explanations: A Unified Framework with Sanity Checks
Tuomas Oikarinen, Ge Yan, Tsui-Wei Weng
Poster Session 2
101 - Truthful or Fabricated? Using Causal Attribution to Mitigate Reward Hacking in Explanations
Pedro Lobato Ferreira, Wilker Aziz, Ivan Titov
Poster Session 2
103 - Identifiable Steering via Sparse Autoencoding of Multi-Concept Shifts
Shruti Joshi, Andrea Dittadi, Sebastien Lachapelle, Dhanya Sridhar
Poster Session 2
109 - Rethinking Crowdsourced Evaluation of Neuron Explanations
Tuomas Oikarinen, Ge Yan, Akshay R. Kulkarni, Tsui-Wei Weng
Poster Session 2
111 - Why Do Some Inputs Break Low-Bit LLMs?
Ting-Yun Chang, Muru Zhang, Jesse Thomason, Robin Jia
Poster Session 2
113 - Are language models aware of the road not taken? Token-level uncertainty and hidden state dynamics
Amir Zur, Eric J Bigelow, Atticus Geiger, Ekdeep Singh Lubana
Poster Session 2
117 - Bayesian Influence Functions for Scalable Data Attribution
Philipp Alexander Kreer, Wilson Wu, Maxwell Adam, Zach Furman, Jesse Hoogland
Poster Session 2
131 - REVIVING YOUR MNEME: Predicting The Side Effects of LLM Unlearning and Fine-Tuning via Sparse Model Diffing
Aly M. Kassem, Golnoosh Farnadi, Negar Rostamzadeh, Zhuan Shi
Poster Session 2
137 - Convergent Linear Representations of Emergent Misalignment
Anna Soligo, Edward Turner, Senthooran Rajamanoharan, Neel Nanda
Poster Session 2
147 - Resilient Multi-Concept Steering in LLMs via Enhanced Sparse "Conditioned" Autoencoders
Saurish Srivastava, Kevin Zhu, Cole Blondin, Sean O'Brien
Poster Session 2
152 - Explanation Design in Strategic Learning: Sufficient Explanations that Induce Non-harmful Responses
Kiet Q. H. Vo, Siu Lun Chau, Masahiro Kato, Yixin Wang, Krikamol Muandet
Poster Session 2
154 - Internal states before wait modulate reasoning patterns
Dmitrii Troitskii, Koyena Pal, Chris Wendler, Callum Stuart McDougall, Neel Nanda
Poster Session 2
Online only
4 - Beyond Black-Box: Structuring Multi-Stage Recommender Systems Using Predicted Intents
Yuyan Wang, Cheenar Banerjee, Samer Chucri, Minmin Chen
Online Only
5 - The Blessing of Reasoning: LLM-Based Contrastive Explanations in Black-Box Recommender Systems
Yuyan Wang, Pan Li, Minmin Chen
Online Only
9 - Steering Self-Evaluation: Interpreting LLM's Reasoning Across Domains and Languages
Praveen Hegde
Online Only
10 - ExpProof : Operationalizing Explanations for Confidential Models with ZKPs
Chhavi Yadav, Evan Laufer, Dan Boneh, Kamalika Chaudhuri
Online Only
19 - Towards Understanding the Mechanisms of Classifier-Free-Guidance
Xiang Li, Rongrong Wang, Qing Qu
Online Only
36 - Persistent Demographic Information in X-ray Foundation Embeddings: a Risk for a Safe and Fair Deployment in Healthcare
Filipe Santos, Aldo Marzullo, Alessandro Quarta, João M. C. Sousa, Susana M. Vieira, Leo Anthony Celi, Francesco Calimeri, Laleh Seyyed-Kalantari
Online Only
41 - What Kind of User Are You? Uncovering User Models in LLM Chatbots
Yida Chen, Aoyu Wu, Trevor DePodesta, Catherine Yeh, Lena Armstrong, Kenneth Li, Nicholas Castillo Marin, Oam Patel, Jan Riecke, Shivam Raval, Olivia Seow, Martin Wattenberg, Fernanda Viégas
Online Only
58 - Breaking Bad Tokens: Detoxification of LLMs Using Sparse Autoencoders
Agam Goyal, Vedant Rathi, William Yeh, Yian Wang, Yuen Chen, Hari Sundaram
Online Only
59 - Learning Augmented Robust Algorithmic Recourse
Kshitij Kayastha, Shahin Jabbari, Vasilis Gkatzelis
Online Only
64 - MPF: Aligning and Debiasing Language Models post Deployment via Multi-Perspective Fusion
Xin Guan, Pei-Hsin Lin, Zekun Wu, Ze Wang, Ruibo Zhang, Emre Kazim, Adriano Koshiyama
Online Only
73 - Joint Localization and Activation Editing for Low-Resource Fine-Tuning
Wen Lai, Alexander Fraser, Ivan Titov
Online Only
106 - BlueGlass: A Framework for Composite AI Safety
Harshal Nandigramwar, Qutub Sha Syed, Kay-Ulrich Scholl
Online Only
108 - Transferring Features Across Language Models With Model Stitching
Alan Chen, Jack Merullo, Alessandro Stolfo, Ellie Pavlick
Online Only
114 - No Training Wheels: Steering Vectors for Bias Correction at Inference Time
Aviral Gupta, Armaan Sethi, Ameesh Sethi
Online Only
133 - Beyond Sparsity: Improving Diversity in Sparse Autoencoders via Denoising Training
Xiang Pan, Yifei Wang, Qi Lei
Online Only
145 - Disentangling and Steering Multilingual Representations: Layer-Wise Analysis and Cross-Lingual Control in Language Models
Abir HARRASSE, Florent Draye, Bernhard Schölkopf, Zhijing Jin
Online Only
148 - FiLMed: Fine-Grained Visual Tokens Align with Localized Semantics
Zhuohao Ni, Xiaoxiao Li
Online Only
149 - On the Effect of Uncertainty on Layer-wise Inference Dynamics
Sunwoo Kim, Haneul Yoo, Alice Oh
Online Only
157 - Why Do Metrics Think That? Towards Understanding Large Language Models as Machine Translation Evaluators
Runzhe Zhan, Xinyi Yang, Junchao Wu, Lidia S. Chao, Derek F. Wong
Online Only
