Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning

• URL: http://arxiv.org/abs/2509.03646v3
• Date: Sat, 27 Sep 2025 17:07:24 GMT
• Title: Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning
• Authors: Haozhe Wang, Qixin Xu, Che Liu, Junhong Wu, Fangzhen Lin, Wenhu Chen,
• Abstract summary: 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.
• Score: 56.496001894673235
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reinforcement Learning (RL) has proven highly effective at enhancing the complex reasoning abilities of Large Language Models (LLMs), yet underlying mechanisms driving this success remain largely opaque. 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, akin to the separation of high-level strategic planning from low-level procedural execution in human cognition. We uncover a compelling two-phase dynamic: initially, a model is constrained by procedural correctness and must improve its low-level skills. The learning bottleneck then decisively shifts, with performance gains being driven by the exploration and mastery of high-level strategic planning. This insight exposes a core inefficiency in prevailing RL algorithms like GRPO, which apply optimization pressure agnostically and dilute the learning signal across all tokens. To address this, we propose Hierarchy-Aware Credit Assignment (HICRA), an algorithm that concentrates optimization efforts on high-impact planning tokens. Our extensive experiments validate that HICRA significantly outperforms strong baselines, and offer deep insights into how reasoning advances through the lens of strategic exploration.

Related papers

• HiMAC: Hierarchical Macro-Micro Learning for Long-Horizon LLM Agents [19.63866851076813]
  HiMAC is a hierarchical agentic RL framework that decomposes long-horizon decision-making into macro-level planning and micro-level execution.<n>Our results show that introducing structured hierarchy, rather than increasing model scale alone, is a key factor for enabling robust long-horizon agentic intelligence.
  arXiv  Detail & Related papers  (2026-03-01T08:09:03Z)
• CARL: Critical Action Focused Reinforcement Learning for Multi-Step Agent [53.56274149236814]
  We propose CARL, a critical-action-focused reinforcement learning algorithm tailored for multi-step agents.<n>Carl achieves both stronger performance and higher efficiency during training and inference across diverse evaluation settings.
  arXiv  Detail & Related papers  (2025-12-04T16:15:46Z)
• Principled RL for Diffusion LLMs Emerges from a Sequence-Level Perspective [85.06838178922791]
  Reinforcement Learning (RL) has proven highly effective for autoregressive language models.<n>But adapting these methods to diffusion large language models (dLLMs) presents fundamental challenges.<n>We propose a principled RL framework that treats entire sequence generation as a single action and uses the ELBO as a tractable sequence-level likelihood proxy.
  arXiv  Detail & Related papers  (2025-12-03T13:05:32Z)
• Demystifying Reinforcement Learning in Agentic Reasoning [90.3737088727791]
  We conduct a comprehensive and systematic investigation to demystify reinforcement learning in agentic reasoning.<n>We highlight our key insights: (i) replacing stitched synthetic trajectories with real end-to-end tool-use trajectories yields a far stronger SFT.<n> Exploration-friendly techniques are crucial for agentic RL, such as clip higher, overlong reward shaping, and maintaining adequate policy entropy could improve the training efficiency.
  arXiv  Detail & Related papers  (2025-10-13T17:57:15Z)
• Decomposing the Entropy-Performance Exchange: The Missing Keys to Unlocking Effective Reinforcement Learning [106.68304931854038]
  Reinforcement learning with verifiable rewards (RLVR) has been widely used for enhancing the reasoning abilities of large language models (LLMs)<n>We conduct a systematic empirical analysis of the entropy-performance exchange mechanism of RLVR across different levels of granularity.<n>Our analysis reveals that, in the rising stage, entropy reduction in negative samples facilitates the learning of effective reasoning patterns.<n>In the plateau stage, learning efficiency strongly correlates with high-entropy tokens present in low-perplexity samples and those located at the end of sequences.
  arXiv  Detail & Related papers  (2025-08-04T10:08:10Z)
• Training Large Language Models to Reason via EM Policy Gradient [0.27195102129094995]
  We introduce an off-policy reinforcement learning algorithm, EM Policy Gradient, to enhance LLM reasoning.<n>We evaluate the effectiveness of EM Policy Gradient on the GSM8K and MATH (HARD) datasets.<n>Models fine-tuned with our method exhibit cognitive behaviors, such as sub-problem decomposition, self-verification, and backtracking.
  arXiv  Detail & Related papers  (2025-04-24T01:31:05Z)
• Extending Token Computation for LLM Reasoning [5.801044612920816]
  Large Language Models (LLMs) are pivotal in advancing natural language processing. LLMs often struggle with complex reasoning tasks due to inefficient attention distributions. We introduce a novel method for extending computed tokens in the Chain-of-Thought process, utilizing attention mechanism optimization.
  arXiv  Detail & Related papers  (2024-03-22T03:23:58Z)
• Entropy-Regularized Token-Level Policy Optimization for Language Agent Reinforcement [67.1393112206885]
  Large Language Models (LLMs) have shown promise as intelligent agents in interactive decision-making tasks. We introduce Entropy-Regularized Token-level Policy Optimization (ETPO), an entropy-augmented RL method tailored for optimizing LLMs at the token level. We assess the effectiveness of ETPO within a simulated environment that models data science code generation as a series of multi-step interactive tasks.
  arXiv  Detail & Related papers  (2024-02-09T07:45:26Z)
• Efficient Reinforcement Learning via Decoupling Exploration and Utilization [6.305976803910899]
  Reinforcement Learning (RL) has achieved remarkable success across multiple fields and applications, including gaming, robotics, and autonomous vehicles. In this work, our aim is to train agent with efficient learning by decoupling exploration and utilization, so that agent can escaping the conundrum of suboptimal Solutions. The above idea is implemented in the proposed OPARL (Optimistic and Pessimistic Actor Reinforcement Learning) algorithm.
  arXiv  Detail & Related papers  (2023-12-26T09:03:23Z)
• PEAR: Primitive Enabled Adaptive Relabeling for Boosting Hierarchical Reinforcement Learning [25.84621883831624]
  Hierarchical reinforcement learning (HRL) has the potential to solve complex long horizon tasks using temporal abstraction and increased exploration.<n>We present primitive enabled adaptive relabeling (PEAR)<n>We first perform adaptive relabeling on a few expert demonstrations to generate efficient subgoal supervision.<n>We then jointly optimize HRL agents by employing reinforcement learning (RL) and imitation learning (IL)
  arXiv  Detail & Related papers  (2023-06-10T09:41:30Z)
• Provable Hierarchy-Based Meta-Reinforcement Learning [50.17896588738377]
  We analyze HRL in the meta-RL setting, where learner learns latent hierarchical structure during meta-training for use in a downstream task. We provide "diversity conditions" which, together with a tractable optimism-based algorithm, guarantee sample-efficient recovery of this natural hierarchy. Our bounds incorporate common notions in HRL literature such as temporal and state/action abstractions, suggesting that our setting and analysis capture important features of HRL in practice.
  arXiv  Detail & Related papers  (2021-10-18T17:56:02Z)

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.