Saved in:
Bibliographic Details
Main Authors: Liu, Siyuan, Yuan, Hongbang, Li, Xinze, Zhu, Ziyue, Cao, Yixin, Jiang, Yu-Gang
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.09503
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918289141137408
author Liu, Siyuan
Yuan, Hongbang
Li, Xinze
Zhu, Ziyue
Cao, Yixin
Jiang, Yu-Gang
author_facet Liu, Siyuan
Yuan, Hongbang
Li, Xinze
Zhu, Ziyue
Cao, Yixin
Jiang, Yu-Gang
contents Large language model (LLM) agents have demonstrated remarkable capabilities in complex decision-making and tool-use tasks, yet their ability to generalize across varying environments remains a under-examined concern. Current evaluation paradigms predominantly rely on trajectory-based metrics that measure task success, while failing to assess whether agents possess a grounded, transferable model of the environment. To address this gap, we propose Task-to-Quiz (T2Q), a deterministic and automated evaluation paradigm designed to decouple task execution from world-state understanding. We instantiate this paradigm in T2QBench, a suite comprising 30 environments and 1,967 grounded QA pairs across multiple difficulty levels. Our extensive experiments reveal that task success is often a poor proxy for environment understanding, and that current memory machanism can not effectively help agents acquire a grounded model of the environment. These findings identify proactive exploration and fine-grained state representation as primary bottlenecks, offering a robust foundation for developing more generalizable autonomous agents.
format Preprint
id arxiv_https___arxiv_org_abs_2601_09503
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle What Do LLM Agents Know About Their World? Task2Quiz: A Paradigm for Studying Environment Understanding
Liu, Siyuan
Yuan, Hongbang
Li, Xinze
Zhu, Ziyue
Cao, Yixin
Jiang, Yu-Gang
Artificial Intelligence
Large language model (LLM) agents have demonstrated remarkable capabilities in complex decision-making and tool-use tasks, yet their ability to generalize across varying environments remains a under-examined concern. Current evaluation paradigms predominantly rely on trajectory-based metrics that measure task success, while failing to assess whether agents possess a grounded, transferable model of the environment. To address this gap, we propose Task-to-Quiz (T2Q), a deterministic and automated evaluation paradigm designed to decouple task execution from world-state understanding. We instantiate this paradigm in T2QBench, a suite comprising 30 environments and 1,967 grounded QA pairs across multiple difficulty levels. Our extensive experiments reveal that task success is often a poor proxy for environment understanding, and that current memory machanism can not effectively help agents acquire a grounded model of the environment. These findings identify proactive exploration and fine-grained state representation as primary bottlenecks, offering a robust foundation for developing more generalizable autonomous agents.
title What Do LLM Agents Know About Their World? Task2Quiz: A Paradigm for Studying Environment Understanding
topic Artificial Intelligence
url https://arxiv.org/abs/2601.09503