NeurIPS 2025 Workshop MechInterp Submissions

Open-Vocabulary Natural-Language Explanations of LLM Activations via Soft Prompts
Bart Bussmann
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Learning to Steer: Input-dependent Steering for Multimodal LLMs
Jayneel Parekh, Pegah KHAYATAN, Mustafa Shukor, Arnaud Dapogny, Alasdair Newson, Matthieu Cord
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Trilemma of Truth in Large Language Models
Germans Savcisens, Tina Eliassi-Rad
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Neurons Speak in Ranges: Breaking Free from Discrete Neuronal Attribution
Muhammad Umair Haider, Hammad Rizwan, Hassan Sajjad, Peizhong Ju, A.B. Siddique
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Finding Manifolds With Bilinear Autoencoders
Thomas Dooms, Ward Gauderis
- Published: 30 Sept 2025, Last Modified: 13 Oct 2025
- Mech Interp Workshop (NeurIPS 2025) Spotlight
- Readers: Everyone
Do Natural Language Descriptions of Model Activations Convey Privileged Information?
Millicent Li, Alberto Mario Ceballos Arroyo, Giordano Rogers, Naomi Saphra, Byron C Wallace
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Watch the Weights: Unsupervised monitoring and control of fine-tuned LLMs
Ziqian Zhong, Aditi Raghunathan
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Fluid Reasoning Representations
Dmitrii Kharlapenko, Alessandro Stolfo, Arthur Conmy, Mrinmaya Sachan, Zhijing Jin
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
WASP: A Weight-Space Approach to Detecting Learned Spuriousness
Cristian Daniel Paduraru, Antonio Barbalau, Radu Filipescu, Andrei Liviu Nicolicioiu, Elena Burceanu
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
False Sense of Security: Why Probing-based Malicious Input Detection Fails to Generalize
Cheng Wang, Zeming Wei, Qin Liu, Wenxuan Zhou, Muhao Chen
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Spotlight
- Readers: Everyone
Rethinking Sparse Autoencoders: Select-and-Project for Fairness and Control from Encoder Features Alone
Antonio Barbalau, Cristian Daniel Paduraru, Teodor Poncu, Alexandru Tifrea, Elena Burceanu
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Adversarial Attacks Leverage Interference Between Features in Superposition
Edward Stevinson, Lucas Prieto, Melih Barsbey, Tolga Birdal
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Spotlight
- Readers: Everyone
Toward a Theory of Generalizability in LLM Mechanistic Interpretability Research
Sean Trott
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning
Helena Casademunt, Caden Juang, Adam Karvonen, Samuel Marks, Senthooran Rajamanoharan, Neel Nanda
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Compressed Computation is (probably) not Computation in Superposition
Jai Bhagat, Sara Molas-Medina, Giorgi Giglemiani, Stefan Heimersheim
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Interpreting learned search: finding a transition model and value function in an RNN that plays Sokoban
Mohammad Taufeeque, Aaron David Tucker, Adam Gleave, Adrià Garriga-Alonso
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Spotlight
- Readers: Everyone
RelP: Faithful and Efficient Circuit Discovery in Language Models via Relevance Patching
Farnoush Rezaei Jafari, Oliver Eberle, Ashkan Khakzar, Neel Nanda
- Published: 30 Sept 2025, Last Modified: 28 Oct 2025
- Mech Interp Workshop (NeurIPS 2025) Spotlight
- Readers: Everyone
Do We Always Need Sampling? Eliciting Numerical Predictive Distributions of LLMs Without Auto-Regression
Julianna Piskorz, Kasia Kobalczyk, Mihaela van der Schaar
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
The Impossibility of Inverse Permutation Learning in Transformer Models
Rohan Alur, Chris Hays, Manish Raghavan, Devavrat Shah
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Interpretability for Time Series Transformers using A Concept Bottleneck Framework
Angela van Sprang, Erman Acar, Willem Zuidema
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
InterpDetect: Interpretable Signals for Detecting Hallucinations in Retrieval-Augmented Generation
Likun Tan, Kuan-Wei Huang, Joy Shi, Kevin Wu
- Published: 30 Sept 2025, Last Modified: 08 Oct 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Some Attention is All You Need for Retrieval
Felix Michalak, Steven Abreu
- Published: 30 Sept 2025, Last Modified: 21 Oct 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Head Pursuit: Probing Attention Specialization in Multimodal Transformers
Lorenzo Basile, Valentino Maiorca, Diego Doimo, Francesco Locatello, Alberto Cazzaniga
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
Mechanistic Interpretability as Statistical Estimation: A Variance Analysis of EAP-IG
Maxime Méloux, François Portet, Maxime Peyrard
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Spotlight
- Readers: Everyone
Mechanistic evidence that motif-gated domain recognition drives contact prediction in protein language models
Jatin Nainani, Bryn Marie Reimer, Connor Watts, David Jensen, Anna G. Green
- Published: 30 Sept 2025, Last Modified: 30 Sept 2025
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone