Evaluating Generative Agents with Actions Grounded in Socially Distributed Task Environments using Incognita

• URL: http://arxiv.org/abs/2607.02975v1
• Date: Fri, 03 Jul 2026 05:34:21 GMT
• Title: Evaluating Generative Agents with Actions Grounded in Socially Distributed Task Environments using Incognita
• Abstract summary: We introduce Incognita, a framework that separates social interaction from grounded execution.<n>We evaluate three generative agent models on 18 tasks stratified by social breadth, with 540 trials.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Effective agency in social environments depends on when an agent seeks knowledge, when it acts, and whether its actions are justified by acquired information. Existing grounded benchmarks provide executable actions, persistent state, and verifiable outcomes, while social simulation environments provide rich interaction among language agents. We study an evaluation setting that combines these requirements. We define socially distributed task environments as interactive environments where task-relevant knowledge is partitioned across role-isolated participants and consequential actions are accessible only through them. Communication serves as exploration over role-partitioned knowledge, while grounded action serves as exploitation over environment state. We introduce Incognita, a Concordia-based framework that separates social interaction from grounded execution. The evaluated agent routes messages to a user or specialist entities; specialists mediate admissible operations; a deterministic sub-environment executes accepted operations over a canonical state; and an offline evaluator scores outcomes with inherited rewards. Incognita-Retail transforms tau-bench retail into a multi-entity environment while preserving final-state reward semantics. We evaluate three generative agent models on 18 tasks stratified by social breadth, with 540 trials. Progress appears in reward and behavior: success rises from 0 percent to 8.9 percent and 17.2 percent, while premature finalization falls from 100 percent to 87 percent and 58 percent. Stronger models elicit more hidden knowledge, contact more entities, and attempt more grounded writes, yet reliability remains low. These findings show that socially distributed task environments expose behavior before reliable success, including knowledge elicitation, source selection, grounded action attempts, and premature completion belief.

Related papers

• Efficient Test-Time Adaptation through Human-AI Interaction [112.5131603022122]
  We propose test-time adaptation through human-agent interaction (TAHI)<n>We adapt agents to 30 individuals in two high-utility domains, writing and visual creation, on a total of 600 tasks.<n>Our agents improve solo task success by 4.5-20.9% within only tens of tasks.
  arXiv  Detail & Related papers  (2026-09-03T17:33:18Z)
• OpenART: Scaling Agent Red Teaming via Open-Ended Environment Evolution [71.06848799741152]
  AI agents operate in persistent environments where early state changes can influence decisions far into the future.<n>Current safety benchmarks often fail to capture cumulative risks because they focus on short, static tasks.<n>We introduce OpenART, an open-ended arena for scalable agent red teaming through environment evolution.
  arXiv  Detail & Related papers  (2026-08-01T13:51:55Z)
• SpatialWorld: Benchmarking Interactive Spatial Reasoning of Multimodal Agents in Real-World Tasks [73.92333717662558]
  SpatialWorld is a benchmark for evaluating the interactive spatial understanding of multimodal agents in complex real-world tasks.<n>It features 760 human-annotated tasks across diverse domains (e.g., household routines, travel, social collaboration)<n>For reliable evaluation, each task includes a human-validated initial state, a reference trajectory, and a terminal-state verifier.
  arXiv  Detail & Related papers  (2026-06-08T15:51:51Z)
• Online Agent-as-a-Judge: Situation-Generating Evaluation for Interactive Agents [7.750851374657493]
  We propose Online Agent-as-a-Judge, a situation-generating evaluation framework for interactive social agents.<n>Online Agent-as-a-Judge deploys an in-world evaluator agent that interacts with the target agent through the environment's native dialogue and action protocol.<n>In a life-simulation environment with $32$ designer-authored social criteria, Online Agent-as-a-Judge improves criteria coverage and agreement with human labels.
  arXiv  Detail & Related papers  (2026-06-06T14:37:33Z)
• Agents Explore but Agents Ignore: LLMs Lack Environmental Curiosity [12.381781997363609]
  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.
  arXiv  Detail & Related papers  (2026-04-19T20:49:41Z)
• Autonomous Continual Learning of Computer-Use Agents for Environment Adaptation [57.65688895630163]
  We introduce ACuRL, an Autonomous Curriculum Reinforcement Learning framework that continually adapts agents to specific environments with zero human data.<n>Our method effectively enables both intra-environment and cross-environment continual learning, yielding 4-22% performance gains without forgetting existing environments.
  arXiv  Detail & Related papers  (2026-02-10T23:06:02Z)
• Automatic Cognitive Task Generation for In-Situ Evaluation of Embodied Agents [43.01384379901339]
  We propose a dynamic in-situ task generation method for unseen environments inspired by human cognition.<n>In the interaction stage, the agent actively interacts with the environment, creating a loop between task execution and generation.<n>Experiments across 10 unseen scenes demonstrate that TEA automatically generated 87,876 tasks in two cycles.
  arXiv  Detail & Related papers  (2026-02-05T03:07:00Z)
• OmniEAR: Benchmarking Agent Reasoning in Embodied Tasks [52.87238755666243]
  We present OmniEAR, a framework for evaluating how language models reason about physical interactions, tool usage, and multi-agent coordination in embodied tasks.<n>We model continuous physical properties and complex spatial relationships across 1,500 scenarios spanning household and industrial domains.<n>Our systematic evaluation reveals severe performance degradation when models must reason from constraints.
  arXiv  Detail & Related papers  (2025-08-07T17:54:15Z)
• Stubborn: An Environment for Evaluating Stubbornness between Agents with Aligned Incentives [4.022057598291766]
  We present Stubborn, an environment for evaluating stubbornness between agents with fully-aligned incentives. In our preliminary results, the agents learn to use their partner's stubbornness as a signal for improving the choices that they make in the environment.
  arXiv  Detail & Related papers  (2023-04-24T17:19:15Z)
• Aligning to Social Norms and Values in Interactive Narratives [89.82264844526333]
  We focus on creating agents that act in alignment with socially beneficial norms and values in interactive narratives or text-based games. We introduce the GALAD agent that uses the social commonsense knowledge present in specially trained language models to contextually restrict its action space to only those actions that are aligned with socially beneficial values.
  arXiv  Detail & Related papers  (2022-05-04T09:54:33Z)

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.