[2603.02479] PRISM: Pushing the Frontier of Deep Think via Process Reward Model-Guided Inference

Computer Science > Artificial Intelligence arXiv:2603.02479 (cs) [Submitted on 3 Mar 2026] Title:PRISM: Pushing the Frontier of Deep Think via Process Reward Model-Guided Inference Authors:Rituraj Sharma, Weiyuan Chen, Noah Provenzano, Tu Vu View a PDF of the paper titled PRISM: Pushing the Frontier of Deep Think via Process Reward Model-Guided Inference, by Rituraj Sharma and 3 other authors View PDF HTML (experimental) Abstract:DEEPTHINK methods improve reasoning by generating, refining, and aggregating populations of candidate solutions, which enables strong performance on complex mathematical and scientific tasks. However, existing frameworks often lack reliable correctness signals during inference, which creates a population-enhancement bottleneck where deeper deliberation amplifies errors, suppresses correct minority solutions, and yields weak returns to additional compute. In this paper, we introduce a functional decomposition of DEEPTHINK systems and propose PRISM, a Process Reward Model (PRM)-guided inference algorithm that uses step-level verification to guide both population refinement and solution aggregation. During refinement, PRISM treats candidate solutions as particles in a PRM-defined energy landscape and reshapes the population through score-guided resampling and stochastic refinement, which concentrates probability mass on higher-quality reasoning while preserving diversity. Across mathematics and science benchmarks, PRISM is competitive with or outperf...