LLM Interpretability

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

57 papers in the last four weeks, up 73% on the four weeks before. 0.6% of all new papers.

Jul 13Week of Sep 28

Latest papers 546

All topics

Interpreting Style Representations via Style-Eliciting Prompts
Jun 4, 2026Junghwan Kim, David JurgensLLM Prompting LLM Interpretability
University of Michigan
A Navigable Manifold of Hypothesized Consciousness-Spectrum States in Language Model Representations
Jun 4, 2026Sophie ZhaoNeural Representation Geometry LLM Interpretability
School of Computer Science Georgia Institute of Technology
Localizing Prompt Ambiguity in Large Language Models with Probe-Targeted Attribution
Jun 3, 2026Govind Ramesh, Yao Dou, Wei XuGradient-Based Attribution LLM Interpretability
Georgia Institute of Technology
Trust, but Don't Verify: Epistemic Blind Spots in LLM Source Evaluation
Jun 3, 2026Rohan N. Pradhan, Steve GoleyLLM Interpretability LLM Reliability
Amazon
Pattern Selectivity is Not Task-Causal Structure: A Cross-Architecture Mechanistic Study of Composed-Task Circuits in 1B-Class Language Models
Jun 3, 2026Yongzhong XuAttention Head Analysis Decoder-Only Language Models
Fast & Faithful Function Vectors
Jun 3, 2026Minh An Pham, Anton Segeler, Thomas Wiegand +4Task Vectors LLM Interpretability
Fraunhofer Heinrich-Hertz-Institute, Berlin, Germany · Technische Universität Berlin, Berlin, Germany · BIFOLD – Berlin Institute for the Foundations of Learning and Data, Berlin, Germany +1
Data Attribution in Large Language Models via Bidirectional Gradient Optimization
Jun 3, 2026Frédéric Berdoz, Luca A. Lanzendörfer, Kaan Bayraktar +1Gradient-Based Attribution LLM Interpretability
ETH Zurich
Do Large Language Models Have Emotions?
Jun 3, 2026Amit Goldenberg, James J. GrossLLM Evaluation LLM Interpretability
Harvard Business School · Department of Psychology, Harvard University · Harvard University, Digital, Data and Design Institute +1
Sparse Mixture-of-Experts Reward Models Learn Interpretable and Specialized Experts for Personalized Preference Modeling
Jun 2, 2026Yifan Wang, Jinyi Mu, Mayank Jobanputra +5Pairwise Preference Learning Reward Modeling
Saarland University · Independent Researcher · Bielefeld University +2
Language Models Compare Quantities Using Number-specific and Unit-specific Heuristics
Jun 2, 2026Mutsumi Sasaki, Go kamoda, Ryosuke Takahashi +4Numerical Reasoning in Language Models LLM Evaluation
Tohoku University · SOKENDAI · NINJAL +2
Framing Migration News with LLMs: Structured CoT as a Support for Human Interpretation
Jun 2, 2026David Alonso del Barrio, Jing Wen, Daniel Gatica-PerezLLM Interpretability CoT Reasoning
Idiap Research Institute and EPFL Switzerland
When Graph Tokens Sink: A Mechanistic Analysis of Graph Language Models
Jun 2, 2026Ding Zhang, Runtao Zhou, Wenqing Zheng +3LLM Interpretability Mechanistic Interpretability
University of Virginia · Capital One
A Close Look At World Model Recovery In Supervised Fine-Tuned LLM Planners
Jun 2, 2026Patrick Emami, Nan Qiang, Peter GrafSupervised Fine-Tuning LLM Planning
National Laboratory of the Rockies · Epic Systems
Decomposing how prompting steers behavior
Jun 2, 2026Fan L. Cheng, Nikolaus KriegeskorteLLM Prompting LLM Interpretability
Columbia University
Multi-component Causal Tracing in Large Language Models
Jun 2, 2026Zirui Yan, Dennis Wei, Dmitriy A. Katz +2LLM Interpretability Causal Reasoning in Language Models
Rensselaer Polytechnic Institute · IBM Research
Perplexity Can Miss SAE Feature Damage Under Quantization
Jun 2, 2026Evan DuanLLM Quantization LLM Interpretability
University of Michigan
Linear Probes Detect Task Format, Not Reasoning Mode in Language Model Hidden States
Jun 1, 2026Subramanyam Sahoo, Vinija Jain, Aman Chadha +1LLM Interpretability Representation Geometry in Language Models
Horizon Research · Meta · Apple +1
On the Persistent Effects of Lexicality in Large Language Models
Jun 1, 2026Hammad Rizwan, Muhammad Umair Haider, Nishant Subramani +3LLM Interpretability Language Modeling
Dalhousie University Halifax, NS · University of Kentucky Lexington, KY · Carnegie Mellon University Pittsburgh, PA
An Enigma of Artificial Reason: Investigating the Production-Evaluation Gap in Large Reasoning Models
May 31, 2026Mingzhong Sun, Teresa Yeo, Armando Solar-Lezama +1Reasoning Evaluation LLM Interpretability
1NUS Department of Computer Science · 4Singapore-MIT Alliance for Research and Technology (SMART) · 2MIT EECS +1
Unlocking the Black Box of Latent Reasoning: An Interpretability-Guided Approach to Intervention
May 31, 2026Shuochen Chang, Tong Bai, Xiaofeng Zhang +5LLM Interpretability Causal Interventions in Language Models
1Shanghai Jiao Tong University · 2Fudan University
The Shape of Wisdom: Decision Trajectories in Language Models
May 31, 2026Shailesh RanaLLM Interpretability
Independent Researcher
The Case for Model Science: Verify, Explore, Steer, Refine
May 31, 2026Przemyslaw Biecek, Luca Longo, Jianlong Zhou +3LLM Evaluation LLM Interpretability
Center for Credible AI · University of Warsaw · Warsaw University of Technology +7
Not All Explanations Simulate Equally: Comparing Verbalized Feature Attributions and Self-Generated Rationales
May 31, 2026Pingjun Hong, Benjamin RothFeature Attribution LLM Interpretability
Faculty of Computer Science, University of Vienna, Vienna, Austria · UniVie Doctoral School Computer Science, University of Vienna, Vienna, Austria · Faculty of Philological and Cultural Studies, University of Vienna, Vienna, Austria
MENTIS: What Belief Changes Under Alignment? Measuring Multi-Scale Latent Torsion in Language Models
May 31, 2026Partha Pratim Saha, Samarth Raina, Mayur Parvatikar +4LLM Interpretability Preference Alignment
Pragya Lab, BITS Pilani Goa, India · IIIT Delhi, India · Amazon, USA +2
Query Lens: Interpreting Sparse Key-Value Features with Indirect Effects
May 30, 2026Hwiyeong Lee, Ingyu Bang, Uiji Hwang +2Representation Learning LLM Interpretability
1Hanyang University, Seoul, Republic of Korea.
How Far Do Auto-Interpretation Labels Generalize: A Controlled Study Across Languages, Scripts, and Rewordings
May 29, 2026Sripad KarneLLM Interpretability Sparse Autoencoders
Columbia University
The Latin Substrate: How Language Models Represent and Mediate Script Choice
May 29, 2026Daniil Gurgurov, Alan Saji, Katharina Trinley +2Multilingual Language Models LLM Interpretability
D German Research Center for AI (DFKI) · I Indian Institute of Technology, Madras · H University of Hamburg +1
Steering LLMs? Actually, Sparse Autoencoders can outperform simple baselines
May 29, 2026Mikkel Godsk Jørgensen, Lars Kai HansenLanguage Model Steering LLM Interpretability
DTU Compute · Technical University of Denmark
The Shape of Addition: Geometric Structures of Arithmetic in Large Language Models
May 29, 2026Liuyuan Wen, Xun Zhu, Lihao Huang +2Neural Representation Geometry LLM Interpretability
State Key Laboratory of Novel Software Technology, Nanjing University, Nanjing 210023, China.
Toxic HallucinAItions: Perturbing Prompts and Tracing LLM Circuits
May 29, 2026Soorya Ram Shimgekar, Agam Goyal, Amruta Parulekar +6Prompt Sensitivity LLM Interpretability
University of Illinois Urbana-Champaign · Nimblemind

Jun 4, 2026, cs.CLJ/K move · Enter open · S save

Interpreting Style Representations via Style-Eliciting Prompts

Junghwan Kim, David Jurgens

University of Michigan

LLM Prompting LLM Interpretability Controllable Text Generation Neural Network Interpretability Text Generation

Open paper AskPDF arXiv

Style representation learning is a powerful tool for authorship analysis and modeling writing style, yet the latent nature of learned representations makes them difficult to interpret. Recent work has attempted to explain these representations by generating natural language descriptions with large language models (LLMs) conditioned on input text. However, such descriptions are often prone to the LLM's biases and hallucinations, and they lack an explicit objective and practical utility. In this work, we propose a novel framework for interpreting style representations through style-eliciting prompts: natural language instructions designed to steer LLMs to generate text that reflects specific stylistic attributes. We curate 1,010 distinct style features spanning 26 stylistic categories and construct a dataset by prompting an LLM to generate text conditioned on these features. Using this data, we train a decoder to generate a style prompt from the style representation of the generated text. We evaluate our approach on three tasks: (1) recovering original style prompts from generated text, (2) generating text in the same style using the recovered prompts, and (3) steering LLM outputs to match the style of human-written texts. Experiments demonstrate that our method consistently outperforms strong baselines that directly prompt LLMs with target text, achieving superior performance in both style description and style imitation. These results highlight that style-eliciting prompts can provide a practical and interpretable interface to stylistic information encoded in style representations.

Momentum

Related topics

Latest papers 546