Agents Explore but Agents Ignore: LLMs Lack Environmental Curiosity

• URL: http://arxiv.org/abs/2604.17609v1
• Date: Sun, 19 Apr 2026 20:49:41 GMT
• Title: Agents Explore but Agents Ignore: LLMs Lack Environmental Curiosity
• Abstract summary: We show that current LLM-based agents struggle to reflect or react to unexpected information.<n>We inject complete task solutions into the agent environments to deliberately expose a task's solution to a model.<n>While agents discover these solutions on Terminal-Bench in 79-81% of runs, they interact, or exploit, them in only 37-50% of cases.
• Score: 12.381781997363609
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: LLM-based agents are assumed to integrate environmental observations into their reasoning: discovering highly relevant but unexpected information should naturally lead to a model exploiting its own discoveries. We show that this assumption is false for current LLM-based agents, which struggle to reflect or react to unexpected information. Across three benchmarks (Terminal-Bench, SWE-Bench, AppWorld), we inject complete task solutions into the agent environments to deliberately expose a task's solution to a model. While agents discover these solutions on Terminal-Bench in 79-81% of runs, they interact, or exploit, them in only 37-50% of cases. This gap is starkest in AppWorld: agents see documentation stating that a command "returns the complete solution to this task" in over 90% of attempts but exploit this in fewer than 7% of trials. We show that agents lack what we call environmental curiosity: the capability to recognize and investigate unexpected but relevant observations in response to environmental stimuli. We identify three main factors influencing environmental curiosity: available tools in the agent scaffold, test-time compute, and training data distribution. Our findings identify configurations that maximize curiosity also achieve the best performance on the unmodified benchmarks. Yet even jointly optimized agents still ignore discovered solutions in the majority of trials: current agents use the environment to fetch expected information, but not to revise their strategy or maximally exploit useful stimuli.

Related papers

• AgentGym2: Benchmarking Large Language Model Agents in De-Idealized Real-World Environments [84.8170304874972]
  We present AgentGym2, a new evaluation framework with task instances grounded in real-world working demands.<n>It measures agents' ability to execute end-to-end procedures, discover tools via exploration, compose tools for unseen tasks, and remain robust to noisy and underspecified information.
  arXiv  Detail & Related papers  (2026-07-06T14:56:51Z)
• AgentEscapeBench: Evaluating Out-of-Domain Tool-Grounded Reasoning in LLM Agents [35.967227261886435]
  AgentEscapeBench tests whether agents can infer, execute, and revise novel tool-use procedures under explicit long-range dependency constraints.<n>Experiments with sixteen LLM agents and human participants show that performance drops sharply as dependency depth increases.<n>Trajectory analysis attributes model failures mainly to breakdowns in long-range state tracking, clue adherence, and intermediate-result propagation.
  arXiv  Detail & Related papers  (2026-05-08T15:59:27Z)
• AJ-Bench: Benchmarking Agent-as-a-Judge for Environment-Aware Evaluation [71.49152943451328]
  We introduce a benchmark AJ-Bench to evaluate Agent-as-a-Judge across three domains-search, data systems, and graphical user interfaces-comprising 155 tasks and 516 annotated trajectories.<n>Experiments demonstrate consistent performance gains over LLM-as-a-Judge baselines, while also revealing substantial open challenges in agent-based verification.
  arXiv  Detail & Related papers  (2026-04-20T13:23:38Z)
• Ask or Assume? Uncertainty-Aware Clarification-Seeking in Coding Agents [4.301199871195023]
  Large Language Model (LLM) agents are increasingly deployed in open-ended domains like software engineering.<n>We evaluate the clarification-seeking abilities of LLM agents on an underspecified variant of SWE-bench Verified.<n>We propose an uncertainty-aware multi-agent scaffold that explicitly decouples underspecification detection from code execution.
  arXiv  Detail & Related papers  (2026-03-27T09:56:26Z)
• What Do LLM Agents Know About Their World? Task2Quiz: A Paradigm for Studying Environment Understanding [50.35012849818872]
  Large language model (LLM) agents have demonstrated remarkable capabilities in complex decision-making and tool-use tasks.<n>We propose Task-to-Quiz (T2Q), a deterministic and automated evaluation paradigm designed to decouple task execution from world-state understanding.<n>Our 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.
  arXiv  Detail & Related papers  (2026-01-14T14:09:11Z)
• Are Your Agents Upward Deceivers? [73.1073084327614]
  Large Language Model (LLM)-based agents are increasingly used as autonomous subordinates that carry out tasks for users.<n>This raises the question of whether they may also engage in deception, similar to how individuals in human organizations lie to superiors to create a good image or avoid punishment.<n>We observe and define agentic upward deception, a phenomenon in which an agent facing environmental constraints conceals its failure and performs actions that were not requested without reporting.
  arXiv  Detail & Related papers  (2025-12-04T14:47:05Z)
• Caution for the Environment: Multimodal LLM Agents are Susceptible to Environmental Distractions [50.5976989558411]
  This paper investigates the faithfulness of multimodal large language model (MLLM) agents in a graphical user interface (GUI) environment.<n>A general scenario is proposed where both the user and the agent are benign, and the environment, while not malicious, contains unrelated content.<n> Experimental results reveal that even the most powerful models, whether generalist agents or specialist GUI agents, are susceptible to distractions.
  arXiv  Detail & Related papers  (2024-08-05T15:16:22Z)
• AgentBench: Evaluating LLMs as Agents [99.12825098528212]
  Large Language Model (LLM) as agents has been widely acknowledged recently.<n>We present AgentBench, a benchmark that consists of 8 distinct environments to assess LLM-as-Agent's reasoning and decision-making abilities.
  arXiv  Detail & Related papers  (2023-08-07T16:08:11Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.