SuperRL: Reinforcement Learning with Supervision to Boost Language Model Reasoning

• URL: http://arxiv.org/abs/2506.01096v2
• Date: Fri, 08 Aug 2025 08:03:03 GMT
• Title: SuperRL: Reinforcement Learning with Supervision to Boost Language Model Reasoning
• Authors: Yihao Liu, Shuocheng Li, Lang Cao, Yuhang Xie, Mengyu Zhou, Haoyu Dong, Xiaojun Ma, Shi Han, Dongmei Zhang,
• Abstract summary: In environments with sparse rewards, reinforcement learning struggles to sample successful trajectories.<n>We introduce SuperRL, a unified training framework that alternates between RL and SFT.<n>Experiments show that SuperRL surpasses vanilla RL by delivering higher sample efficiency, stronger generalization, and improved robustness under sparse rewards.
• Score: 42.54530036364341
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models are increasingly used for complex reasoning tasks where high-quality offline data such as expert-annotated solutions and distilled reasoning traces are often available. However, in environments with sparse rewards, reinforcement learning struggles to sample successful trajectories, leading to inefficient learning. At the same time, these offline trajectories that represent correct reasoning paths are not utilized by standard on-policy reinforcement learning methods. We introduce SuperRL, a unified training framework that adaptively alternates between RL and SFT. Whenever every rollout for a given instance receives zero reward, indicating the absence of a learning signal, SuperRL falls back to SFT on the curated offline data. Extensive experiments across diverse reasoning benchmarks show that SuperRL surpasses vanilla RL by delivering higher sample efficiency, stronger generalization, and improved robustness under sparse rewards.

Related papers

• Learn Hard Problems During RL with Reference Guided Fine-tuning [56.56461712665904]
  Reinforcement learning (RL) for mathematical reasoning can suffer from reward sparsity.<n>We introduce Reference-Guided Fine-Tuning (ReGFT) to synthesize positive trajectories on hard problems and train on them before RL.<n>Our results show that ReGFT effectively overcomes reward sparsity and unlocks stronger RL-based mathematical reasoning.
  arXiv  Detail & Related papers  (2026-03-01T18:41:28Z)
• Supervised Reinforcement Learning: From Expert Trajectories to Step-wise Reasoning [49.22815446849924]
  Large Language Models (LLMs) often struggle with problems that require multi-step reasoning.<n>For small-scale open-source models, Reinforcement Learning with Verifiable Rewards (RLVR) fails when correct solutions are rarely sampled.<n>We propose Supervised Reinforcement Learning (SRL), a framework that reformulates problem solving as generating a sequence of logical "actions"
  arXiv  Detail & Related papers  (2025-10-29T22:05:08Z)
• RoiRL: Efficient, Self-Supervised Reasoning with Offline Iterative Reinforcement Learning [4.311472216447055]
  Reinforcement learning is central to improving reasoning in large language models (LLMs)<n>We propose RoiRL: Reasoning with offline iterative Reinforcement Learning.<n>We show that RoiRL trains to 2.5x faster and consistently outperforms TTRL on reasoning benchmarks.
  arXiv  Detail & Related papers  (2025-10-03T10:59:26Z)
• 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)
• ConfClip: Confidence-Weighted and Clipped Reward for Reinforcement Learning in LLMs [32.13266235550995]
  Reinforcement learning (RL) has become a standard paradigm for refining large language models (LLMs)<n>Inspired by observations from human learning, we introduce a RL technique that integrates verifiable outcomes with the model's own confidence estimates.
  arXiv  Detail & Related papers  (2025-09-22T13:00:35Z)
• Scaling Up RL: Unlocking Diverse Reasoning in LLMs via Prolonged Training [121.5858973157225]
  We investigate the effects of prolonged reinforcement learning on a small language model across a diverse set of reasoning domains.<n>We introduce controlled KL regularization, clipping ratio, and periodic reference policy resets as critical components for unlocking long-term performance gains.<n>Our model achieves significant improvements over strong baselines, including +14.7% on math, +13.9% on coding, and +54.8% on logic puzzle tasks.
  arXiv  Detail & Related papers  (2025-07-16T17:59:24Z)
• Consistent Paths Lead to Truth: Self-Rewarding Reinforcement Learning for LLM Reasoning [87.7836502955847]
  We propose a novel self-rewarding reinforcement learning framework to enhance Large Language Model (LLM) reasoning.<n>Our key insight is that correct responses often exhibit consistent trajectory patterns in terms of model likelihood.<n>We introduce CoVo, an intrinsic reward mechanism that integrates Consistency and Volatility via a robust vector-space aggregation strategy.
  arXiv  Detail & Related papers  (2025-06-10T12:40:39Z)
• 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)
• Diffusion Guidance Is a Controllable Policy Improvement Operator [98.11511661904618]
  CFGRL is trained with the simplicity of supervised learning, yet can further improve on the policies in the data.<n>On offline RL tasks, we observe a reliable trend -- increased guidance weighting leads to increased performance.
  arXiv  Detail & Related papers  (2025-05-29T14:06:50Z)
• SeRL: Self-Play Reinforcement Learning for Large Language Models with Limited Data [65.56911325914582]
  We propose Self-play Reinforcement Learning (SeRL) to bootstrap Large Language Models (LLMs) training with limited initial data.<n>The proposed SeRL yields results superior to its counterparts and achieves performance on par with those obtained by high-quality data with verifiable rewards.
  arXiv  Detail & Related papers  (2025-05-25T13:28:04Z)
• Large Language Model-Enhanced Reinforcement Learning for Generic Bus Holding Control Strategies [12.599164162404994]
  This study introduces an automatic reward generation paradigm by leveraging the in-context learning and reasoning capabilities of Large Language Models (LLMs)<n>To evaluate the feasibility of the proposed LLM-enhanced RL paradigm, it is applied to extensive bus holding control scenarios that vary in the number of bus lines, stops, and passenger demand.
  arXiv  Detail & Related papers  (2024-10-14T07:10:16Z)
• Hybrid Inverse Reinforcement Learning [34.793570631021005]
  inverse reinforcement learning approach to imitation learning is a double-edged sword. We propose using hybrid RL -- training on a mixture of online and expert data -- to curtail unnecessary exploration. We derive both model-free and model-based hybrid inverse RL algorithms with strong policy performance guarantees.
  arXiv  Detail & Related papers  (2024-02-13T23:29:09Z)
• SMORE: Score Models for Offline Goal-Conditioned Reinforcement Learning [33.125187822259186]
  Offline Goal-Conditioned Reinforcement Learning (GCRL) is tasked with learning to achieve multiple goals in an environment purely from offline datasets using sparse reward functions. We present a novel approach to GCRL under a new lens of mixture-distribution matching, leading to our discriminator-free method: SMORe.
  arXiv  Detail & Related papers  (2023-11-03T16:19:33Z)
• Mastering the Unsupervised Reinforcement Learning Benchmark from Pixels [112.63440666617494]
  Reinforcement learning algorithms can succeed but require large amounts of interactions between the agent and the environment. We propose a new method to solve it, using unsupervised model-based RL, for pre-training the agent. We show robust performance on the Real-Word RL benchmark, hinting at resiliency to environment perturbations during adaptation.
  arXiv  Detail & Related papers  (2022-09-24T14:22:29Z)
• Improving Zero-shot Generalization in Offline Reinforcement Learning using Generalized Similarity Functions [34.843526573355746]
  Reinforcement learning (RL) agents are widely used for solving complex sequential decision making tasks, but exhibit difficulty in generalizing to scenarios not seen during training. We show that performance of online algorithms for generalization in RL can be hindered in the offline setting due to poor estimation of similarity between observations. We propose a new theoretically-motivated framework called Generalized Similarity Functions (GSF), which uses contrastive learning to train an offline RL agent to aggregate observations based on the similarity of their expected future behavior.
  arXiv  Detail & Related papers  (2021-11-29T15:42:54Z)
• Offline Meta-Reinforcement Learning with Online Self-Supervision [66.42016534065276]
  We propose a hybrid offline meta-RL algorithm, which uses offline data with rewards to meta-train an adaptive policy. Our method uses the offline data to learn the distribution of reward functions, which is then sampled to self-supervise reward labels for the additional online data. We find that using additional data and self-generated rewards significantly improves an agent's ability to generalize.
  arXiv  Detail & Related papers  (2021-07-08T17:01:32Z)
• Text Generation with Efficient (Soft) Q-Learning [91.47743595382758]
  Reinforcement learning (RL) offers a more flexible solution by allowing users to plug in arbitrary task metrics as reward. We introduce a new RL formulation for text generation from the soft Q-learning perspective. We apply the approach to a wide range of tasks, including learning from noisy/negative examples, adversarial attacks, and prompt generation.
  arXiv  Detail & Related papers  (2021-06-14T18:48:40Z)

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.