# Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning

arXiv: https://arxiv.org/abs/2509.03646 · HTML: https://mlsanity.com/paper/2509.03646/

## Key facts

- Cited by 42 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning for reasoning, one of 17 directions in LLM reasoning.
- Builds on 30 of its 44 references that are in the map.
- First posted on arXiv 3 Sep 2025.

## Placed in

- LLM reasoning / [Reinforcement learning for reasoning](https://mlsanity.com/lr/reinforcement-learning-for-reasoning/)

## Builds on (most cited first, 30 of 30)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [DAPO: An Open-Source LLM Reinforcement Learning System at Scale](https://mlsanity.com/paper/2503.14476/) (2025)
- [MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts](https://mlsanity.com/paper/2310.02255/) (2023)
- [Understanding R1-Zero-Like Training: A Critical Perspective](https://mlsanity.com/paper/2503.20783/) (2025)
- [OpenAI o1 System Card](https://mlsanity.com/paper/2412.16720/) (2024)
- [Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://mlsanity.com/paper/2501.12599/) (2025)
- [OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems](https://mlsanity.com/paper/2402.14008/) (2024)
- [Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning](https://mlsanity.com/paper/2506.01939/) (2025)
- [The Entropy Mechanism of Reinforcement Learning for Reasoning Language Models](https://mlsanity.com/paper/2505.22617/) (2025)
- [Reinforcement Learning and Control as Probabilistic Inference: Tutorial and Review](https://mlsanity.com/paper/1805.00909/) (2018)
- [VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning](https://mlsanity.com/paper/2504.08837/) (2025)
- [Reasoning with Exploration: An Entropy Perspective](https://mlsanity.com/paper/2506.14758/) (2025)
- [Pixel Reasoner: Incentivizing Pixel-Space Reasoning with Curiosity-Driven Reinforcement Learning](https://mlsanity.com/paper/2505.15966/) (2025)
- [Hierarchical Reasoning Model](https://mlsanity.com/paper/2506.21734/) (2025)
- [Can MLLMs Reason in Multimodality? EMMA: An Enhanced MultiModal ReAsoning Benchmark](https://mlsanity.com/paper/2501.05444/) (2025)
- [MiMo-VL Technical Report](https://mlsanity.com/paper/2506.03569/) (2025)
- [SEED-GRPO: Semantic Entropy Enhanced GRPO for Uncertainty-Aware Policy Optimization](https://mlsanity.com/paper/2505.12346/) (2025)
- [Beyond Distillation: Pushing the Limits of Medical LLM Reasoning with Minimalist Rule-Based RL](https://mlsanity.com/paper/2505.17952/) (2025)
- [To Code or not to Code? Adaptive Tool Integration for Math Language Models via Expectation-Maximization](https://mlsanity.com/paper/2502.00691/) (2025)
- [Learning to Reason via Mixture-of-Thought for Logical Reasoning](https://mlsanity.com/paper/2505.15817/) (2025)
- [HS-STaR: Hierarchical Sampling for Self-Taught Reasoners via Difficulty Estimation and Budget Reallocation](https://mlsanity.com/paper/2505.19866/) (2025)
- [R1-RE: Cross-Domain Relation Extraction with RLVR](https://mlsanity.com/paper/2507.04642/) (2025)
- [Afford-X: Generalizable and Slim Affordance Reasoning for Task-oriented Manipulation](https://mlsanity.com/paper/2503.03556/) (2025)

## Built on by (most cited first, 30 of 42)

- [ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory](https://mlsanity.com/paper/2509.25140/) (2025)
- [A Survey of Reinforcement Learning for Large Reasoning Models](https://mlsanity.com/paper/2509.08827/) (2025)
- [Agent0: Unleashing Self-Evolving Agents from Zero Data via Tool-Integrated Reasoning](https://mlsanity.com/paper/2511.16043/) (2025)
- [Rethinking Entropy Interventions in RLVR: An Entropy Change Perspective](https://mlsanity.com/paper/2510.10150/) (2025)
- [Embarrassingly Simple Self-Distillation Improves Code Generation](https://mlsanity.com/paper/2604.01193/) (2026)
- [NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents](https://mlsanity.com/paper/2510.07172/) (2025)
- [RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMs](https://mlsanity.com/paper/2509.21128/) (2025)
- [Demystifying Reinforcement Learning in Agentic Reasoning](https://mlsanity.com/paper/2510.11701/) (2025)
- [Reverse-Engineered Reasoning for Open-Ended Generation](https://mlsanity.com/paper/2509.06160/) (2025)
- [From Reasoning to Agentic: Credit Assignment in Reinforcement Learning for Large Language Models](https://mlsanity.com/paper/2604.09459/) (2026)
- [The Debate on RLVR Reasoning Capability Boundary: Shrinkage, Expansion, or Both? A Two-Stage Dynamic View](https://mlsanity.com/paper/2510.04028/) (2025)
- [On the Direction of RLVR Updates for LLM Reasoning: Identification and Exploitation](https://mlsanity.com/paper/2603.22117/) (2026)
- [From Illusion to Intention: Visual Rationale Learning for Vision-Language Reasoning](https://mlsanity.com/paper/2511.23031/) (2025)
- [Step Potential Advantage Estimation: Harnessing Intermediate Confidence and Correctness for Efficient Mathematical Reasoning](https://mlsanity.com/paper/2601.03823/) (2026)
- [From Narrow to Panoramic Vision: Attention-Guided Cold-Start Reshapes Multimodal Reasoning](https://mlsanity.com/paper/2603.03825/) (2026)
- [Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation](https://mlsanity.com/paper/2607.05382/) (2026)
- [Bad Seeing or Bad Thinking? Rewarding Perception for Multimodal Reasoning](https://mlsanity.com/paper/2605.14054/) (2026)
- [Training Large Language Models To Reason In Parallel With Global Forking Tokens](https://mlsanity.com/paper/2510.05132/) (2025)
- [Towards a Mechanistic Understanding of Large Reasoning Models: A Survey of Training, Inference, and Failures](https://mlsanity.com/paper/2601.19928/) (2026)
- [ReLaX: Reasoning with Latent Exploration for Large Reasoning Models](https://mlsanity.com/paper/2512.07558/) (2025)
- [Beyond Uniform Credit Assignment: Selective Eligibility Traces for RLVR](https://mlsanity.com/paper/2605.05965/) (2026)
- [ReasoningFlow: Discourse Structures for Understanding LLM Reasoning Traces](https://mlsanity.com/paper/2606.05402/) (2026)
- [CogDoc: Towards Unified thinking in Documents](https://mlsanity.com/paper/2512.12658/) (2025)
- [G2TR: Generation-Guided Visual Token Reduction for Separate-Encoder Unified Multimodal Models](https://mlsanity.com/paper/2605.12309/) (2026)
- [Explore-Execute Chain: Towards an Efficient Structured Reasoning Paradigm](https://mlsanity.com/paper/2509.23946/) (2025)
- [Selection, Reflection and Self-Refinement: Revisit Reasoning Tasks via a Causal Lens](https://mlsanity.com/paper/2510.08222/) (2025)
- [The Fragility of Strategic Thinking in Large Language Models](https://mlsanity.com/paper/2510.10813/) (2025)
- [Measuring and Mitigating Post-hoc Rationalization in Reverse Chain-of-Thought Generation](https://mlsanity.com/paper/2602.14469/) (2026)
- [RL Post-Training Builds Compositional Reasoning Strategies](https://mlsanity.com/paper/2607.07646/) (2026)
- [Language-Driven Object-Oriented Two-Stage Method for Scene Graph Anticipation](https://mlsanity.com/paper/2509.05661/) (2025)

## Abstract (arXiv)

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.