TriEx: A Game-based Tri-View Framework for Explaining Internal Reasoning in Multi-Agent LLMs
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913053757407232 |
|---|---|
| author | Wang, Ziyi Zhang, Chen Peng, Wenjun Wu, Qi Wang, Xinyu |
| author_facet | Wang, Ziyi Zhang, Chen Peng, Wenjun Wu, Qi Wang, Xinyu |
| contents | Explainability for Large Language Model (LLM) agents is especially challenging in interactive, partially observable settings, where decisions depend on evolving beliefs and other agents. We present \textbf{TriEx}, a tri-view explainability framework that instruments sequential decision making with aligned artifacts: (i) structured first-person self-reasoning bound to an action, (ii) explicit second-person belief states about opponents updated over time, and (iii) third-person oracle audits grounded in environment-derived reference signals. This design turns explanations from free-form narratives into evidence-anchored objects that can be compared and checked across time and perspectives. Using imperfect-information strategic games as a controlled testbed, we show that TriEx enables scalable analysis of explanation faithfulness, belief dynamics, and evaluator reliability, revealing systematic mismatches between what agents say, what they believe, and what they do. Our results highlight explainability as an interaction-dependent property and motivate multi-view, evidence-grounded evaluation for LLM agents. Code is available at https://github.com/Einsam1819/TriEx. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_20043 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | TriEx: A Game-based Tri-View Framework for Explaining Internal Reasoning in Multi-Agent LLMs Wang, Ziyi Zhang, Chen Peng, Wenjun Wu, Qi Wang, Xinyu Computation and Language Artificial Intelligence Explainability for Large Language Model (LLM) agents is especially challenging in interactive, partially observable settings, where decisions depend on evolving beliefs and other agents. We present \textbf{TriEx}, a tri-view explainability framework that instruments sequential decision making with aligned artifacts: (i) structured first-person self-reasoning bound to an action, (ii) explicit second-person belief states about opponents updated over time, and (iii) third-person oracle audits grounded in environment-derived reference signals. This design turns explanations from free-form narratives into evidence-anchored objects that can be compared and checked across time and perspectives. Using imperfect-information strategic games as a controlled testbed, we show that TriEx enables scalable analysis of explanation faithfulness, belief dynamics, and evaluator reliability, revealing systematic mismatches between what agents say, what they believe, and what they do. Our results highlight explainability as an interaction-dependent property and motivate multi-view, evidence-grounded evaluation for LLM agents. Code is available at https://github.com/Einsam1819/TriEx. |
| title | TriEx: A Game-based Tri-View Framework for Explaining Internal Reasoning in Multi-Agent LLMs |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2604.20043 |