Not All Prompts Are Equal: Exploration-Guided Prompt Scaffolding for Multimodal Reinforcement Post-Training

• URL: http://arxiv.org/abs/2609.15051v1
• Date: Mon, 14 Sep 2026 05:12:05 GMT
• Title: Not All Prompts Are Equal: Exploration-Guided Prompt Scaffolding for Multimodal Reinforcement Post-Training
• Abstract summary: Training prompts in online reinforcement learning (RL) differ substantially in how informative they are for the current policy.<n>We propose an exploration-guided prompt scaffolding framework that adapts the training prompt distribution dynamically throughout RL post-training of multimodal large language models (MLLMs)<n>We use a teacher model to generate scaffolded rewrites that preserve the original task intent while making subsequent training more informative.
• Score: 15.647521419163985
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Training prompts in online reinforcement learning (RL) differ substantially in how informative they are for the current policy: some are already saturated while others are too difficult to yield reliable learning signals, yet both receive equal rollout budget under standard training. We propose an exploration-guided prompt scaffolding framework that adapts the training prompt distribution dynamically throughout RL post-training of multimodal large language models (MLLMs). Central to our approach is the $\textit{Exploration Potential Score} (EPS)$, a lightweight rollout-based proxy for prompt utility derived from KL-regularized policy improvement theory, computable directly from on-policy rollout statistics without additional overhead. Rather than discarding low-utility prompts, we use a teacher model to generate scaffolded rewrites that preserve the original task intent while making subsequent training more informative, reframing teacher supervision as training-data refinement rather than output imitation. Integrated with GRPO on Geo3K and MMK12, our method consistently outperforms the baseline on both in-domain and out-of-distribution benchmarks, achieving up to 9.7\% relative improvement in-domain and gains of 11.5\% on MathVision and 11.1\% on MMMU-Pro.

Related papers

• LLMs Are Already Good Tutors: Training-Free Prompt Optimization for Pedagogical Math Tutoring [1.4976621159818981]
  We investigate whether training-free prompt optimization-evolving only the system prompt via API calls can serve as a practical alternative.<n>We adapt 7 published methods and propose 5 education-specialized methods, evaluating these 12 methods under 5 conditions on 2 OOD benchmark suites.<n> Behavioral analysis with an 82-code educational codebook reveals that training-free methods rely on teaching-knowledge patterns at 2-3x the rate of RL-trained models.
  arXiv  Detail & Related papers  (2026-05-26T14:35:57Z)
• Spend Your Rollouts Where It Counts: Rollout Allocation for Group-Based RL Post-Training [34.81218483901967]
  Reinforcement learning is the dominant paradigm for post-training large language models.<n>Group-based policy optimization methods compute advantages from multiple rollouts per prompt.<n>Pilot-Commit is a budget-aware rollout allocation framework for group-based RL post-training.
  arXiv  Detail & Related papers  (2026-05-26T06:41:13Z)
• MM-Doc-R1: Training Agents for Long Document Visual Question Answering through Multi-turn Reinforcement Learning [74.07254720088926]
  We introduce MM-Doc-R1, a novel framework that employs an agentic, vision-aware workflow to address long document visual question answering.<n>We propose Similarity-based Policy Optimization (SPO), addressing baseline estimation bias in existing multi-turn reinforcement learning (RL) algorithms like GRPO.<n>Our experiments on the MMLongbench-Doc benchmark show that MM-Doc-R1 outperforms previous baselines by 10.4%.
  arXiv  Detail & Related papers  (2026-04-15T07:39:08Z)
• SOUP: Token-level Single-sample Mix-policy Reinforcement Learning for Large Language Models [67.41779761651924]
  SOUP is a framework that unifies off- and on-policy learning within individual samples at the token level.<n>It consistently outperforms standard on-policy training and existing off-policy extensions.
  arXiv  Detail & Related papers  (2026-01-29T09:56:15Z)
• Stronger Together: On-Policy Reinforcement Learning for Collaborative LLMs [20.084201133669534]
  Multi-agent systems (MAS) and reinforcement learning (RL) are widely used to enhance the agentic capabilities of large language models (LLMs)<n>Applying on-policy RL to MAS remains underexplored and presents unique challenges.<n>We propose AT-GRPO, which includes (i) an agent- and turn-wise grouped RL algorithm tailored to MAS and (ii) a training system that supports both single- and multi-policy regimes.
  arXiv  Detail & Related papers  (2025-10-13T06:55:09Z)
• DiFFPO: Training Diffusion LLMs to Reason Fast and Furious via Reinforcement Learning [37.20873499361773]
  We propose a unified framework for training masked diffusion large language models (dLLMs) to reason better (furious)<n>We first unify the existing baseline approach by proposing to train surrogate policies via off-policy RL, whose likelihood is much more tractable as an approximation to the true dLLM policy.<n>We also propose a new direction of joint training efficient samplers/controllers of dLLMs policy. Via RL, we incentivize dLLMs' natural multi-token prediction capabilities by letting the model learn to adaptively allocate an inference threshold for each prompt.
  arXiv  Detail & Related papers  (2025-10-02T16:57:24Z)
• Agentic Reinforcement Learning with Implicit Step Rewards [92.26560379363492]
  Large language models (LLMs) are increasingly developed as autonomous agents using reinforcement learning (agentic RL)<n>We introduce implicit step rewards for agentic RL (iStar), a general credit-assignment strategy that integrates seamlessly with standard RL algorithms.<n>We evaluate our method on three challenging agent benchmarks, including WebShop and VisualSokoban, as well as open-ended social interactions with unverifiable rewards in SOTOPIA.
  arXiv  Detail & Related papers  (2025-09-23T16:15:42Z)
• Your Reward Function for RL is Your Best PRM for Search: Unifying RL and Search-Based TTS [62.22644307952087]
  We introduce AIRL-S, the first natural unification of RL-based and search-based TTS.<n>We leverage adversarial inverse reinforcement learning (AIRL) combined with group relative policy optimization (GRPO) to learn a dense, dynamic PRM directly from correct reasoning traces.<n>Our results show that our unified approach improves performance by 9 % on average over the base model, matching GPT-4o.
  arXiv  Detail & Related papers  (2025-08-19T23:41:15Z)
• DUMP: Automated Distribution-Level Curriculum Learning for RL-based LLM Post-training [25.91869315787235]
  We present a principled curriculum learning framework grounded in the notion of distribution-level learnability.<n>Our framework prioritizes distributions with either high average advantage (exploitation) or low sample count (exploration)<n>Our experiments show that our framework significantly improves convergence speed and final performance.
  arXiv  Detail & Related papers  (2025-04-13T20:10:27Z)
• Fast Adaptation with Behavioral Foundation Models [82.34700481726951]
  Unsupervised zero-shot reinforcement learning has emerged as a powerful paradigm for pretraining behavioral foundation models.<n>Despite promising results, zero-shot policies are often suboptimal due to errors induced by the unsupervised training process.<n>We propose fast adaptation strategies that search in the low-dimensional task-embedding space of the pre-trained BFM to rapidly improve the performance of its zero-shot policies.
  arXiv  Detail & Related papers  (2025-04-10T16:14:17Z)
• VinePPO: Refining Credit Assignment in RL Training of LLMs [66.80143024475635]
  We propose VinePPO, a straightforward approach that leverages the flexibility of language environments to compute unbiased Monte Carlo-based estimates.<n>Our method consistently outperforms PPO and other baselines across MATH and GSM8K datasets in less wall-clock time.
  arXiv  Detail & Related papers  (2024-10-02T15:49:30Z)
• MUSBO: Model-based Uncertainty Regularized and Sample Efficient Batch Optimization for Deployment Constrained Reinforcement Learning [108.79676336281211]
  Continuous deployment of new policies for data collection and online learning is either cost ineffective or impractical. We propose a new algorithmic learning framework called Model-based Uncertainty regularized and Sample Efficient Batch Optimization. Our framework discovers novel and high quality samples for each deployment to enable efficient data collection.
  arXiv  Detail & Related papers  (2021-02-23T01:30:55Z)

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.