The First Workshop on the Application of LLM Explainability to Reasoning and Planning
@ COLM 2025
We are thrilled to announce the First Workshop on the Application of LLM Explainability to Reasoning and Planning at COLM 2025 to be held on October 10, 2025.
Enabling large language models (LLMs) to reason (e.g., arithmetic reasoning, symbolic reasoning, commonsense reasoning, etc.) and plan (e.g., path-finding, tool use, web navigation, computer use, etc.) has been a popular topic in the past few years. Despite the exciting achievement, there have also been growing concerns about the safety and trustworthiness of these LLM applications, due to our large “unknowns” on how LLMs achieve these capabilities and where they could fail. On the other hand, LLM explainability (broadly including any research explaining or interpreting LLMs) has also attracted increasing attention, but existing research has mostly focused on simplified tasks and hardly yields insights that can be directly applied to realistic reasoning and planning tasks. This discrepancy has consequently raised doubts about the practical meaning of LLM explainability research.
In this workshop, we aim to bring together researchers from various perspectives to discuss the potential and practical applications of model explainability to advancing LLM reasoning and planning. Specifically, the workshop welcomes submissions on the following topics (non-exclusively):
Join our Google Group for workshop updates and Q&A https://groups.google.com/g/xllm-reasoning-planning-workshop, and contact us at xllmreasoningplanningworkshop AT gmail DOT com for other inquiries!
| The Geometry of Self-Verification in a Task-Specific Reasoning Model Andrew Lee, Lihao Sun, Chris Wendler, Fernanda Viégas, Martin Wattenberg |
| Attributing Response to Context: A Jensen–Shannon Divergence Driven Mechanistic Study of Context Attribution in Retrieval-Augmented Generation Ruizhe Li, Chen Chen, Yuchen Hu, Yanjun Gao, Xi Wang, Emine Yilmaz |
| When Models Know More Than They Can Explain: Quantifying Knowledge Transfer in Human-AI Collaboration Quan Shi, Carlos E Jimenez, Shunyu Yao, Nick Haber, Diyi Yang, Karthik R Narasimhan |
| Enhancing Logical Reasoning in Large Language Models through Graph-based Synthetic Data Jiaming Zhou, Abbas Ghaddar, Ge Zhang, Liheng Ma, Yaochen Hu, Soumyasundar Pal, Bin Wang, Jianye HAO, Mark Coates, Yingxue Zhang |
| Disambiguate First, Parse Later: Generating Interpretations for Ambiguity Resolution in Semantic Parsing Irina Saparina, Mirella Lapata |
| Beyond Autocomplete: Designing CopilotLens Towards Transparent and Explainable AI Coding Agents Runlong Ye, Zeling Zhang, Boushra Almazroua, Michael Liut |
| Latent Chain-of-Thought? Decoding the Depth-Recurrent Transformer Wenquan Lu, Yuechuan Yang, Kyle Lee, Yanshu Li, Enqi Liu |
| Rethinking (Human) Preference Evaluation of LLM Rationales Ziang Li, Manasi Ganti, Zixian Ma, Helena Vasconcelos, Qijia He, Ranjay Krishna |
| HYBRIDMIND: Meta Selection of Natural Language and Symbolic Language for Enhanced LLM Reasoning Simeng Han, Tianyu Liu, Chuhan Li, Xuyuan Xiong, Arman Cohan |
| Angular Steering: Behavior Control via Rotation in Activation Space Hieu M. Vu, Tan Minh Nguyen |
| Can LLMs Reason Abstractly Over Math Word Problems Without CoT? Disentangling Abstract Formulation From Arithmetic Computation Ziling Cheng, Meng Cao, Leila Pishdad, Yanshuai Cao, Jackie CK Cheung |
| Creativity or Brute Force? Using Brainteasers as a Window into the Problem-Solving Abilities of Large Language Models Simeng Han, Stephen Xia, Grant Zhang, Howard Dai, Chen Liu, Lichang Chen, Hoang H Nguyen, Hongyuan Mei, Jiayuan Mao, R. Thomas McCoy |
| Case-Based Reasoning Enhances the Predictive Power of LLMs in Drug-Drug Interaction Guangyi Liu, Yongqi Zhang, Xunyuan Liu, Quanming Yao |
| Are General-Purpose LLMs Ready for Planning? A Large-Scale Evaluation in PDDL Kaustubh Vyas, Damien Graux, Sebastien Montella, Pavlos Vougiouklis, Jeff Z. Pan |
| ReCalibrate: RL for Uncertainty-Aware Reasoning in LLMs Mehul Damani, Isha Puri, Stewart Slocum, Idan Shenfeld, Jacob Andreas |
| Failure by Interference: Language Models Make Balanced Parentheses Errors When Faulty Mechanisms Overshadow Sound Ones Daking Rai, Samuel Miller, Kevin Moran, Ziyu Yao |
| From Indirect Object Identification to Syllogisms: Exploring Binary Mechanisms in Transformer Circuits Karim Saraipour, Shichang Zhang |
| How Post-Training Reshapes LLMs: A Mechanistic View on Knowledge, Truthfulness, Refusal, and Confidence Hongzhe Du, Weikai Li, Min Cai, Karim Saraipour, Zimin Zhang, Yizhou Sun, Himabindu Lakkaraju, Shichang Zhang |
| Everything is Plausible: Investigating the Impact of LLM Rationales on Human Notions of Plausibility Shramay Palta, Peter A. Rankel, Sarah Wiegreffe, Rachel Rudinger |
| Before You 〈think/〉, Monitor: Implementing Flavell's Metacognitive Framework in LLMs Nick Oh |
| Reasoning Riddles: How Explainability Reveals Cognitive Limits in Vision-Language Models Prahitha Movva |
We welcome both long (up to 9 pages of main content, plus unlimited references) and short (up to 5 pages of main content, plus unlimited references) paper submissions, following the official template of COLM. The long papers are expected to include completed and full-scope work while the short paper submissions can be preliminary or ongoing work. All submissions will be non-archival. We also allow dual submissions that are under review or have recently been accepted to other venues—for the former, authors should make sure to follow the dual submission policies from the other venue; for the latter, we ask authors to indicate the accepted venue.
The workshop will announce one Best Paper Award targeting all authors, and one Special Recognition Award targeting papers with junior and/or underrepresented-group authors being the first authors. Authors submitting to our workshop will be requested to clarify the status of the first author(s) for eligibility confirmation.
We sincerely thank the program committee for their considered and thoughtful reviews!
