Learning to Reason as Action Abstractions with Scalable Mid-Training RL

• URL: http://arxiv.org/abs/2509.25810v1
• Date: Tue, 30 Sep 2025 05:34:20 GMT
• Title: Learning to Reason as Action Abstractions with Scalable Mid-Training RL
• Authors: Shenao Zhang, Donghan Yu, Yihao Feng, Bowen Jin, Zhaoran Wang, John Peebles, Zirui Wang,
• Abstract summary: An effective mid-training phase should identify a compact set of useful actions and enable fast selection.<n>We propose Reasoning as Action Abstractions (RA3), a scalable mid-training algorithm.
• Score: 55.24192942739207
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models excel with reinforcement learning (RL), but fully unlocking this potential requires a mid-training stage. An effective mid-training phase should identify a compact set of useful actions and enable fast selection among them through online RL. We formalize this intuition by presenting the first theoretical result on how mid-training shapes post-training: it characterizes an action subspace that minimizes both the value approximation error from pruning and the RL error during subsequent planning. Our analysis reveals two key determinants of mid-training effectiveness: pruning efficiency, which shapes the prior of the initial RL policy, and its impact on RL convergence, which governs the extent to which that policy can be improved via online interactions. These results suggest that mid-training is most effective when the decision space is compact and the effective horizon is short, highlighting the importance of operating in the space of action abstractions rather than primitive actions. Building on these insights, we propose Reasoning as Action Abstractions (RA3), a scalable mid-training algorithm. Specifically, we derive a sequential variational lower bound and optimize it by iteratively discovering temporally-consistent latent structures via RL, followed by fine-tuning on the bootstrapped data. Experiments on code generation tasks demonstrate the effectiveness of our approach. Across multiple base models, RA3 improves the average performance on HumanEval and MBPP by 8 and 4 points over the base model and the next-token prediction baseline. Furthermore, RA3 achieves faster convergence and higher asymptotic performance in RLVR on HumanEval+, MBPP+, LiveCodeBench, and Codeforces.

Related papers

• On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models [73.10315509190623]
  Recent reinforcement learning techniques have yielded impressive reasoning improvements in language models.<n>It remains unclear whether post-training truly extends a model's reasoning ability beyond what it acquires during pre-training.<n>We develop a fully controlled experimental framework that isolates the causal contributions of pre-training, mid-training, and RL-based post-training.
  arXiv  Detail & Related papers  (2025-12-08T18:12:10Z)
• From Supervision to Exploration: What Does Protein Language Model Learn During Reinforcement Learning? [76.288870982181]
  Protein language models (PLMs) have advanced computational protein science through large-scale pretraining and scalable architectures.<n> reinforcement learning (RL) has broadened exploration and enabled precise multi-objective optimization in protein design.<n>We ask if RL improves sampling efficiency and, more importantly, if it reveals capabilities not captured by supervised learning.
  arXiv  Detail & Related papers  (2025-10-02T01:31:10Z)
• Reinforcement Learning on Pre-Training Data [55.570379963147424]
  We introduce Reinforcement Learning on Pre-Training data (R), a new training-time scaling paradigm for optimizing large language models (LLMs)<n>R enables the policy to autonomously explore meaningful trajectories to learn from pre-training data and improve its capability through reinforcement learning (RL)<n>Extensive experiments on both general-domain and mathematical reasoning benchmarks across multiple models validate the effectiveness of R.
  arXiv  Detail & Related papers  (2025-09-23T17:10:40Z)
• Shuffle-R1: Efficient RL framework for Multimodal Large Language Models via Data-centric Dynamic Shuffle [53.239242017802056]
  Reinforcement learning (RL) has emerged as an effective post-training paradigm for enhancing the reasoning capabilities of multimodal large language model (MLLM)<n>However, current RL pipelines often suffer from training inefficiencies caused by two underexplored issues: Advantage Collapsing and Rollout Silencing.<n>We propose Shuffle-R1, a simple yet principled framework that improves RL fine-tuning efficiency by dynamically restructuring trajectory sampling and batch composition.
  arXiv  Detail & Related papers  (2025-08-07T17:53:47Z)
• Ring-lite: Scalable Reasoning via C3PO-Stabilized Reinforcement Learning for LLMs [51.21041884010009]
  Ring-lite is a Mixture-of-Experts (MoE)-based large language model optimized via reinforcement learning (RL)<n>Our approach matches the performance of state-of-the-art (SOTA) small-scale reasoning models on challenging benchmarks.
  arXiv  Detail & Related papers  (2025-06-17T17:12:34Z)
• Beyond Accuracy: Dissecting Mathematical Reasoning for LLMs Under Reinforcement Learning [93.00629872970364]
  Reinforcement learning (RL) has become the dominant paradigm for improving the performance of language models on complex reasoning tasks.<n>We introduce SPARKLE, a fine-grained analytic framework to dissect the effects of RL across three key dimensions.<n>We study whether difficult problems -- those yielding no RL signals and mixed-quality reasoning traces -- can still be effectively used for training.
  arXiv  Detail & Related papers  (2025-06-05T07:53:59Z)
• A Snapshot of Influence: A Local Data Attribution Framework for Online Reinforcement Learning [37.62558445850573]
  We propose an algorithm, iterative influence-based filtering (IIF), for online RL training.<n>IIF reduces sample complexity, speeds up training, and achieves higher returns.<n>These results advance interpretability, efficiency, and effectiveness of online RL.
  arXiv  Detail & Related papers  (2025-05-25T19:25:57Z)
• DreamerV3 for Traffic Signal Control: Hyperparameter Tuning and Performance [4.962905815955427]
  Reinforcement learning (RL) has evolved into a widely investigated technology for the development of smart TSC strategies.<n>The DreamerV3 algorithm presents compelling properties for policy learning.<n>In this paper, a corridor TSC model is trained using the DreamerV3 algorithm to explore the benefits of world models for TSC strategy learning.
  arXiv  Detail & Related papers  (2025-03-04T05:02:46Z)
• POAR: Efficient Policy Optimization via Online Abstract State Representation Learning [6.171331561029968]
  State Representation Learning (SRL) is proposed to specifically learn to encode task-relevant features from complex sensory data into low-dimensional states. We introduce a new SRL prior called domain resemblance to leverage expert demonstration to improve SRL interpretations. We empirically verify POAR to efficiently handle tasks in high dimensions and facilitate training real-life robots directly from scratch.
  arXiv  Detail & Related papers  (2021-09-17T16:52:03Z)

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.