MAP: A Map-then-Act Paradigm for Long-Horizon Interactive Agent Reasoning

Hugging Face Daily Papers Papers

Summary

The paper proposes the Map-then-Act Paradigm (MAP), a plug-and-play framework that shifts environmental understanding before execution in interactive LLM agents, achieving consistent gains across benchmarks and enabling frontier models to surpass near-zero baseline performance in 22 of 25 game environments.

Current interactive LLM agents rely on goal-conditioned stepwise planning, where environmental understanding is acquired reactively during execution rather than established beforehand. This temporal inversion leads to Delayed Environmental Perception: agents must infer environmental constraints through trial-and-error, resulting in an Epistemic Bottleneck that traps them in inefficient failure cycles. Inspired by human affordance perception and cognitive map theory, we propose the Map-then-Act Paradigm (MAP), a plug-and-play framework that shifts environment understanding before execution. MAP consists of three stages: (1) Global Exploration, acquiring environment-general priors; (2) Task-Specific Mapping, constructing a structured cognitive map; and (3) Knowledge-Augmented Execution, solving tasks grounded on the map. Experiments show consistent gains across benchmarks and LLMs. On ARC-AGI-3, MAP enables frontier models to surpass near-zero baseline performance in 22 of 25 game environments. We further introduce MAP-2K, a dataset of map-then-act trajectories, and show that training on it outperforms expert execution traces, suggesting that understanding environments is more fundamental than imitation.
Original Article
View Cached Full Text

Cached at: 05/14/26, 04:16 AM

Paper page - MAP: A Map-then-Act Paradigm for Long-Horizon Interactive Agent Reasoning

Source: https://huggingface.co/papers/2605.13037

Abstract

Interactive LLM agents suffer from delayed environmental perception and epistemic bottlenecks due to reactive understanding during execution, which the proposed Map-then-Act Paradigm (MAP) addresses by acquiring environmental knowledge beforehand through global exploration, task-specific mapping, and knowledge-augmented execution.

Current interactive LLM agents rely ongoal-conditioned stepwise planning, whereenvironmental understandingis acquired reactively during execution rather than established beforehand. This temporal inversion leads to Delayed Environmental Perception: agents must infer environmental constraints throughtrial-and-error, resulting in anEpistemic Bottleneckthat traps them in inefficient failure cycles. Inspired by humanaffordance perceptionandcognitive map theory, we propose theMap-then-Act Paradigm(MAP), a plug-and-play framework that shifts environment understanding before execution. MAP consists of three stages: (1)Global Exploration, acquiring environment-general priors; (2)Task-Specific Mapping, constructing a structured cognitive map; and (3)Knowledge-Augmented Execution, solving tasks grounded on the map. Experiments show consistent gains across benchmarks and LLMs. OnARC-AGI-3, MAP enables frontier models to surpass near-zero baseline performance in 22 of 25 game environments. We further introduceMAP-2K, a dataset of map-then-act trajectories, and show that training on it outperforms expert execution traces, suggesting that understanding environments is more fundamental than imitation.

View arXiv pageView PDFAdd to collection

Get this paper in your agent:

hf papers read 2605\.13037

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2605.13037 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2605.13037 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2605.13037 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

SpatialAct: Probing Spatial Reasoning-to-Action Capabilities of VLM Agents in 3D Scenes

Hugging Face Daily Papers

SpatialAct is a new simulator-grounded benchmark that probes whether VLM agents can perform coherent spatial reasoning and translate it into actions in 3D environments across multi-turn feedback settings. Experiments reveal a significant reasoning-to-action gap, with current VLMs struggling to maintain spatial beliefs and produce reliable actions despite performing well on isolated reasoning tasks.

From Trainee to Trainer: LLM-Designed Training Environment for RL with Multi-Agent Reasoning

arXiv cs.CL

This paper proposes the LLM-as-Environment-Engineer framework, where a policy model analyzes failures to automatically redesign the training environment for reinforcement learning, and introduces MAPF-FrozenLake as a controllable testbed. The framework, using Qwen3-4B, outperforms larger models like GPT and Gemini, showing that policy learning improves the model's ability to diagnose weaknesses.