SCI-PRM: A Tool Aware Process Reward Model for Scientific Reasoning Verification

• URL: http://arxiv.org/abs/2606.04579v1
• Date: Wed, 03 Jun 2026 08:13:27 GMT
• Title: SCI-PRM: A Tool Aware Process Reward Model for Scientific Reasoning Verification
• Abstract summary: We build a large-scale dataset featuring Chain-of-Tool trajectories that explicitly interleave reasoning with the execution of scientific tools.<n>We then train an efficient reward model called Sci-PRM to provide fine-grained supervision on tool selection, execution accuracy, and result interpretation at each step in one inference.
• Score: 47.20109789089479
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: While Process Reward Models (PRMs) have achieved remarkable success in mathematical reasoning, their application in complex scientific domains-such as biology, chemistry, and physics remains largely unexplored. Scientific problems demand not only logical rigor but also factual consistency and the precise usage of domain-specific tools, areas where current models often suffer from hallucinations and lack of verification. In this paper, we first construct SCIPRM70K, a large-scale dataset featuring Chain-of-Tool trajectories that explicitly interleave reasoning with the execution of scientific tools. Building upon this, we train an efficient reward model called Sci-PRM to provide fine-grained supervision on tool selection, execution accuracy, and result interpretation at each step in one inference. Experiments demonstrate that Sci-PRM significantly enhances foundation models in two key aspects: (1) it enables effective test-time scaling via Best-of-N selection; and (2) when integrated into Reinforcement Learning, it serves as a dense reward signal that mitigates the critical issue of advantage disappearance, allowing the model to break through existing performance ceilings.

Related papers

• MOT-SR: Multi-Objective Tool-Augmented Scientific Equation Discovery with Large Language Models [23.532834629603798]
  Symbolic Regression (SR) aims to discover analytical equations from observational data.<n>We propose Multi-Objective Tool-augmented Symbolic Regression (MOT-SR)<n>MOT-SR integrates external analytical tools to extract structural priors and guide equation generation.
  arXiv  Detail & Related papers  (2026-07-31T15:52:07Z)
• Faithful, Enriched, and Precise: Benchmarking Natural-Science Illustration Generation by T2I models [49.93401412148884]
  We introduce FEPBench, a benchmark built from carefully selected high-quality scientific illustrations.<n>We evaluate T2I models along three dimensions: instruction faithfulness, reasoning enrichment, and semantic precision.<n>Results show that even state-of-the-art closed-source models, such as GPT Image 2 and Nano Banana Pro, still suffer from text-rendering bottlenecks, limited reasoning enrichment, and difficulty balancing generation richness with precision.
  arXiv  Detail & Related papers  (2026-06-04T09:49:02Z)
• Grounding LLMs in Scientific Discovery via Embodied Actions [84.11877211907647]
  Large Language Models (LLMs) have shown significant potential in scientific discovery but struggle to bridge the gap between theoretical reasoning and physical simulation.<n>We propose EmbodiedAct, a framework that transforms established scientific software into active embodied agents by groundings in embodied actions with a tight perception-execution loop.
  arXiv  Detail & Related papers  (2026-02-24T07:37:18Z)
• Sci-CoE: Co-evolving Scientific Reasoning LLMs via Geometric Consensus with Sparse Supervision [15.806243963561776]
  Sci-CoE is a two-stage scientific co-evolving framework that enables models to self-evolve as both solver and verifier.<n>In the first stage, the model uses a small set of annotated data to establish correctness judgment anchors for the Verifier.<n>In the second stage, we introduce a geometric reward mechanism that jointly considers consensus, reliability, and diversity, driving large-scale self-iteration.
  arXiv  Detail & Related papers  (2026-02-12T16:46:00Z)
• Dual-Phase LLM Reasoning: Self-Evolved Mathematical Frameworks [48.105258051884384]
  This paper proposes a new two-stage training framework that enhances models' self-correction capabilities.<n>During the first stage, a multi-turn dialogue strategy guides the model to generate long chain-of-thought (CoT) data.<n>The second stage employs a difficulty-aware rejection sampling mechanism to dynamically optimize data distribution.
  arXiv  Detail & Related papers  (2026-01-09T08:19:11Z)
• CoSineVerifier: Tool-Augmented Answer Verification for Computation-Oriented Scientific Questions [32.14674040685995]
  We introduce model, a tool-augmented verifier that leverages external rubrics to perform precise computations and symbolic simplifications.<n>Experiments conducted on STEM subjects, general QA, and long-form reasoning tasks demonstrates strong generalization of model.
  arXiv  Detail & Related papers  (2025-12-01T03:08:43Z)
• NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents [65.85967483058705]
  Large language models are emerging as powerful tools for scientific law discovery.<n>Existing benchmarks for this task suffer from a fundamental methodological trilemma.<n>We introduce NewtonBench, a benchmark comprising 324 scientific law discovery tasks across 12 physics domains.
  arXiv  Detail & Related papers  (2025-10-08T16:12:11Z)
• Sycophancy Mitigation Through Reinforcement Learning with Uncertainty-Aware Adaptive Reasoning Trajectories [58.988535279557546]
  We introduce textbf sycophancy Mitigation through Adaptive Reasoning Trajectories.<n>We show that SMART significantly reduces sycophantic behavior while preserving strong performance on out-of-distribution inputs.
  arXiv  Detail & Related papers  (2025-09-20T17:09:14Z)
• Improving Molecular Modeling with Geometric GNNs: an Empirical Study [56.52346265722167]
  This paper focuses on the impact of different canonicalization methods, (2) graph creation strategies, and (3) auxiliary tasks, on performance, scalability and symmetry enforcement. Our findings and insights aim to guide researchers in selecting optimal modeling components for molecular modeling tasks.
  arXiv  Detail & Related papers  (2024-07-11T09:04:12Z)
• Let's Reinforce Step by Step [10.65244642965387]
  We use Reinforcement Learning from Human Feedback to shape model reasoning processes. Our results show that the fine-grained reward provided by PRM-based methods enhances accuracy on simple mathematical reasoning. We also show the critical role reward aggregation functions play in model performance.
  arXiv  Detail & Related papers  (2023-11-10T01:35:51Z)

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.