Think Fast and Slow: Step-Level Cognitive Depth Adaptation for LLM Agents

• URL: http://arxiv.org/abs/2602.12662v1
• Date: Fri, 13 Feb 2026 06:52:09 GMT
• Title: Think Fast and Slow: Step-Level Cognitive Depth Adaptation for LLM Agents
• Authors: Ruihan Yang, Fanghua Ye, Xiang We, Ruoqing Zhao, Kang Luo, Xinbo Xu, Bo Zhao, Ruotian Ma, Shanyi Wang, Zhaopeng Tu, Xiaolong Li, Deqing Yang, Linus,
• Abstract summary: Large language models (LLMs) are increasingly deployed as autonomous agents for multi-turn decision-making tasks.<n>This paper introduces Cog, a framework that trains agents to dynamically adapt cognitive depth at each step.<n> Experiments on ALFWorld and ScienceWorld demonstrate that Cog achieves state-of-the-art performance with superior efficiency.
• Score: 49.119608399413806
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) are increasingly deployed as autonomous agents for multi-turn decision-making tasks. However, current agents typically rely on fixed cognitive patterns: non-thinking models generate immediate responses, while thinking models engage in deep reasoning uniformly. This rigidity is inefficient for long-horizon tasks, where cognitive demands vary significantly from step to step, with some requiring strategic planning and others only routine execution. In this paper, we introduce CogRouter, a framework that trains agents to dynamically adapt cognitive depth at each step. Grounded in ACT-R theory, we design four hierarchical cognitive levels ranging from instinctive responses to strategic planning. Our two-stage training approach includes Cognition-aware Supervised Fine-tuning (CoSFT) to instill stable level-specific patterns, and Cognition-aware Policy Optimization (CoPO) for step-level credit assignment via confidence-aware advantage reweighting. The key insight is that appropriate cognitive depth should maximize the confidence of the resulting action. Experiments on ALFWorld and ScienceWorld demonstrate that CogRouter achieves state-of-the-art performance with superior efficiency. With Qwen2.5-7B, it reaches an 82.3% success rate, outperforming GPT-4o (+40.3%), OpenAI-o3 (+18.3%), and GRPO (+14.0%), while using 62% fewer tokens.

Related papers

• Reflecting with Two Voices: A Co-Adaptive Dual-Strategy Framework for LLM-Based Agent Decision Making [24.534365665776672]
  Large language model (LLM) agents often rely on external demonstrations or retrieval-augmented planning.<n>We propose DuSAR - a demonstration-free framework that enables a single frozen LLM to perform co-adaptive reasoning.<n>On ALFWorld and Mind2Web, DuSAR achieves state-of-the-art performance with open-source LLMs.
  arXiv  Detail & Related papers  (2025-12-09T08:44:59Z)
• Towards a Science of Scaling Agent Systems [79.64446272302287]
  We formalize a definition for agent evaluation and characterize scaling laws as the interplay between agent quantity, coordination structure, modelic, and task properties.<n>We derive a predictive model using coordination metrics, that cross-validated R2=0, enabling prediction on unseen task domains.<n>We identify three effects: (1) a tool-coordination trade-off: under fixed computational budgets, tool-heavy tasks suffer disproportionately from multi-agent overhead, and (2) a capability saturation: coordination yields diminishing or negative returns once single-agent baselines exceed 45%.
  arXiv  Detail & Related papers  (2025-12-09T06:52:21Z)
• Multi-Agent Reinforcement Learning and Real-Time Decision-Making in Robotic Soccer for Virtual Environments [0.0]
  This paper presents a unified Multi-Agent Reinforcement Learning (MARL) framework that addresses these challenges.<n>To ensure scalability, we integrate mean-field theory into the HRL framework.<n>Our mean-field actor-critic method achieves a significant performance boost.
  arXiv  Detail & Related papers  (2025-12-02T19:11:44Z)
• Mobile-Agent-RAG: Driving Smart Multi-Agent Coordination with Contextual Knowledge Empowerment for Long-Horizon Mobile Automation [57.12284831164602]
  Mobile agents show immense potential, yet current state-of-the-art (SoTA) agents exhibit inadequate success rates on real-world, long-horizon, cross-application tasks.<n>We propose Mobile-Agent-RAG, a novel hierarchical multi-agent framework that innovatively integrates dual-level retrieval augmentation.
  arXiv  Detail & Related papers  (2025-11-15T15:22:42Z)
• Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning [56.496001894673235]
  Reinforcement Learning (RL) has proven highly effective at enhancing the complex reasoning abilities of Large Language Models (LLMs)<n>Our analysis reveals that puzzling phenomena like aha moments", length-scaling'' and entropy dynamics are not disparate occurrences but hallmarks of an emergent reasoning hierarchy.
  arXiv  Detail & Related papers  (2025-09-03T18:52:49Z)
• KAT-V1: Kwai-AutoThink Technical Report [50.84483585850113]
  We present Kwaipilot-AutoThink (KAT), an open-source 40B large language model developed to address the overthinking problem in reasoning-intensive tasks.<n>KAT dynamically switches between reasoning and non-reasoning modes based on task complexity.<n>We also propose Step-SRPO, a reinforcement learning algorithm that incorporates intermediate supervision into the GRPO framework.
  arXiv  Detail & Related papers  (2025-07-11T04:07:10Z)
• Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents [44.34340798542]
  Large Language Models (LLMs) have shown remarkable capabilities in natural language tasks requiring complex reasoning. Traditional supervised pre-training on static datasets falls short in enabling autonomous agent capabilities. We propose a framework that combines guided Monte Carlo Tree Search (MCTS) search with a self-critique mechanism and iterative fine-tuning on agent interactions.
  arXiv  Detail & Related papers  (2024-08-13T20:52:13Z)
• Ladder-of-Thought: Using Knowledge as Steps to Elevate Stance Detection [73.31406286956535]
  We introduce the Ladder-of-Thought (LoT) for the stance detection task. LoT directs the small LMs to assimilate high-quality external knowledge, refining the intermediate rationales produced. Our empirical evaluations underscore LoT's efficacy, marking a 16% improvement over GPT-3.5 and a 10% enhancement compared to GPT-3.5 with CoT on stance detection task.
  arXiv  Detail & Related papers  (2023-08-31T14:31:48Z)

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.