From Outcomes to Processes: Guiding PRM Learning from ORM for Inference-Time Alignment

• URL: http://arxiv.org/abs/2506.12446v2
• Date: Sat, 28 Jun 2025 14:16:58 GMT
• Title: From Outcomes to Processes: Guiding PRM Learning from ORM for Inference-Time Alignment
• Abstract summary: We introduce process reward models (PRMs) into reward-guided search (RGS)<n>We propose SP-PRM, a novel dual-consistency framework integrating score consistency-based and preference-based partial evaluation modules without relying on human annotation.<n>Experiments on dialogue, summarization, and reasoning tasks demonstrate that SP-PRM substantially enhances existing RGS methods, achieving a 3.6%-10.3% improvement in GPT-4 evaluation scores across all tasks.
• Score: 23.463402040567615
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Inference-time alignment methods have gained significant attention for their efficiency and effectiveness in aligning large language models (LLMs) with human preferences. However, existing dominant approaches using reward-guided search (RGS) primarily rely on outcome reward models (ORMs), which suffer from a critical granularity mismatch: ORMs are designed to provide outcome rewards for complete responses, while RGS methods rely on process rewards to guide the policy, leading to inconsistent scoring and suboptimal alignment. To address this challenge, we introduce process reward models (PRMs) into RGS and argue that an ideal PRM should satisfy two objectives: Score Consistency, ensuring coherent evaluation across partial and complete responses, and Preference Consistency, aligning partial sequence assessments with human preferences. Based on these, we propose SP-PRM, a novel dual-consistency framework integrating score consistency-based and preference consistency-based partial evaluation modules without relying on human annotation. Extensive experiments on dialogue, summarization, and reasoning tasks demonstrate that SP-PRM substantially enhances existing RGS methods, achieving a 3.6%-10.3% improvement in GPT-4 evaluation scores across all tasks.

Related papers

• PAPO: Stabilizing Rubric Integration Training via Decoupled Advantage Normalization [31.935482701047544]
  We propose Process-Aware Policy Optimization (PAPO) to address two limitations of existing reward designs.<n>PAPO integrates process-level evaluation into Group Relative Policy Optimization (GRPO) through decoupled advantage normalization.<n> Experiments across multiple model scales and six benchmarks demonstrate that PAPO consistently outperforms ORM.
  arXiv  Detail & Related papers  (2026-03-27T15:48:13Z)
• Evaluate-as-Action: Self-Evaluated Process Rewards for Retrieval-Augmented Agents [24.080021799509847]
  textscEvalAct (Evaluate-as-Action) converts implicit retrieval quality assessment into an explicit action.<n>textscEvalAct achieves the best average accuracy, with the largest gains on multi-hop tasks.
  arXiv  Detail & Related papers  (2026-03-10T05:22:40Z)
• CPO: Addressing Reward Ambiguity in Role-playing Dialogue via Comparative Policy Optimization [53.79487826635141]
  Reinforcement Learning Fine-Tuning (RLFT) has achieved notable success in tasks with objectively verifiable answers.<n>But it struggles with open-ended subjective tasks like role-playing dialogue.<n>Traditional reward modeling approaches, which rely on independent sample-wise scoring, face dual challenges: subjective evaluation criteria and unstable reward signals.<n>Motivated by the insight that human evaluation inherently combines explicit criteria with implicit comparative judgments, we propose Comparative Policy Optimization.
  arXiv  Detail & Related papers  (2025-08-12T16:49:18Z)
• The Bidirectional Process Reward Model [9.082060895625958]
  We propose a novel bidirectional evaluation paradigm, named Bidirectional Process Reward Model (BiPRM)<n>BiPRM seamlessly incorporates a parallel right-to-left (R2L) evaluation stream alongside the conventional L2R flow, enabling later reasoning steps to help assess earlier ones in real time.<n>We conduct extensive experiments on two mathematical reasoning benchmarks using samples generated by three different policy models.
  arXiv  Detail & Related papers  (2025-08-03T09:23:49Z)
• Dynamic and Generalizable Process Reward Modeling [74.36829922727026]
  We propose Dynamic and Generalizable Process Reward Modeling (DG-PRM), which features a reward tree to capture and store fine-grained, multi-dimensional reward criteria.<n> Experimental results show that DG-PRM achieves stunning performance on prevailing benchmarks, significantly boosting model performance across tasks with dense rewards.
  arXiv  Detail & Related papers  (2025-07-23T18:17:22Z)
• SPARE: Single-Pass Annotation with Reference-Guided Evaluation for Automatic Process Supervision and Reward Modelling [58.05959902776133]
  We introduce Single-Pass.<n>with Reference-Guided Evaluation (SPARE), a novel structured framework that enables efficient per-step annotation.<n>We demonstrate SPARE's effectiveness across four diverse datasets spanning mathematical reasoning (GSM8K, MATH), multi-hop question answering (MuSiQue-Ans), and spatial reasoning (SpaRP)<n>On ProcessBench, SPARE demonstrates data-efficient out-of-distribution generalization, using only $sim$16% of training samples compared to human-labeled and other synthetically trained baselines.
  arXiv  Detail & Related papers  (2025-06-18T14:37:59Z)
• RAG-Zeval: Towards Robust and Interpretable Evaluation on RAG Responses through End-to-End Rule-Guided Reasoning [64.46921169261852]
  RAG-Zeval is a novel end-to-end framework that formulates faithfulness and correctness evaluation as a rule-guided reasoning task.<n>Our approach trains evaluators with reinforcement learning, facilitating compact models to generate comprehensive and sound assessments.<n>Experiments demonstrate RAG-Zeval's superior performance, achieving the strongest correlation with human judgments.
  arXiv  Detail & Related papers  (2025-05-28T14:55:33Z)
• On the Role of Feedback in Test-Time Scaling of Agentic AI Workflows [71.92083784393418]
  Agentic AI (systems that autonomously plan and act) are becoming widespread, yet their task success rate on complex tasks remains low.<n>Inference-time alignment relies on three components: sampling, evaluation, and feedback.<n>We introduce Iterative Agent Decoding (IAD), a procedure that repeatedly inserts feedback extracted from different forms of critiques.
  arXiv  Detail & Related papers  (2025-04-02T17:40:47Z)
• R-PRM: Reasoning-Driven Process Reward Modeling [53.06844294668382]
  Process Reward Models (PRMs) have emerged as a promising solution by evaluating each reasoning step.<n>Existing PRMs typically output evaluation scores directly, limiting both learning efficiency and evaluation accuracy.<n>We propose Reasoning-Driven Process Reward Modeling (R-PRM)<n>R-PRM generates seed data from limited annotations, effectively bootstrapping our model's reasoning capabilities.
  arXiv  Detail & Related papers  (2025-03-27T09:23:08Z)
• MPBench: A Comprehensive Multimodal Reasoning Benchmark for Process Errors Identification [27.594868471770475]
  Reasoning is an essential capacity for large language models (LLMs) to address complex tasks.<n>Process-level reward models (PRMs) were proposed to provide step-wise rewards that facilitate reinforcement learning and data production.<n>Existing benchmarks of PRMs are text-based and focus on error detection, neglecting other scenarios like reasoning search.<n>MPBench is a comprehensive, multi-task, multimodal benchmark designed to systematically assess the effectiveness of PRMs in diverse scenarios.
  arXiv  Detail & Related papers  (2025-03-16T13:50:38Z)
• UMB@PerAnsSumm 2025: Enhancing Perspective-Aware Summarization with Prompt Optimization and Supervised Fine-Tuning [8.095763327154335]
  We present our approach to the PerAnsSumm Shared Task, which involves perspective span identification and perspective-aware summarization.<n>For span identification, we adopt ensemble learning that integrates three transformer models through averaging to exploit individual model strengths.<n>For summarization, we design a suite of Chain-of-Thought (CoT) prompting strategies that incorporate keyphrases and guide information to structure summary generation into manageable steps.
  arXiv  Detail & Related papers  (2025-03-14T06:29:51Z)
• ReARTeR: Retrieval-Augmented Reasoning with Trustworthy Process Rewarding [25.329712997545794]
  We propose Retrieval-Augmented Reasoning through Trustworthy Process Rewarding (ReARTeR)<n>ReARTeR enhances RAG systems' reasoning capabilities through post-training and test-time scaling.<n> Experimental results on multi-step reasoning benchmarks demonstrate significant improvements.
  arXiv  Detail & Related papers  (2025-01-14T05:56:26Z)
• The Lessons of Developing Process Reward Models in Mathematical Reasoning [62.165534879284735]
  Process Reward Models (PRMs) aim to identify and mitigate intermediate errors in the reasoning processes.<n>We develop a consensus filtering mechanism that effectively integrates Monte Carlo (MC) estimation with Large Language Models (LLMs)<n>We release a new state-of-the-art PRM that outperforms existing open-source alternatives.
  arXiv  Detail & Related papers  (2025-01-13T13:10:16Z)
• Revisiting Reciprocal Recommender Systems: Metrics, Formulation, and Method [60.364834418531366]
  We propose five new evaluation metrics that comprehensively and accurately assess the performance of RRS. We formulate the RRS from a causal perspective, formulating recommendations as bilateral interventions. We introduce a reranking strategy to maximize matching outcomes, as measured by the proposed metrics.
  arXiv  Detail & Related papers  (2024-08-19T07:21:02Z)
• How Can I Improve? Using GPT to Highlight the Desired and Undesired Parts of Open-ended Responses [11.809647985607935]
  We explore a sequence labeling approach focused on identifying components of desired and less desired praise for providing explanatory feedback. To quantify the quality of highlighted praise components identified by GPT models, we introduced a Modified Intersection over Union (M-IoU) score. Our findings demonstrate that: (1) the M-IoU score effectively correlates with human judgment in evaluating sequence quality; (2) using two-shot prompting on GPT-3.5 resulted in decent performance in recognizing effort-based and outcome-based praise; and (3) our optimally fine-tuned GPT-3.5 model achieved M-IoU scores of 0.6
  arXiv  Detail & Related papers  (2024-05-01T02:59:10Z)
• Choosing the Best of Both Worlds: Diverse and Novel Recommendations through Multi-Objective Reinforcement Learning [68.45370492516531]
  We introduce Scalarized Multi-Objective Reinforcement Learning (SMORL) for the Recommender Systems (RS) setting. SMORL agent augments standard recommendation models with additional RL layers that enforce it to simultaneously satisfy three principal objectives: accuracy, diversity, and novelty of recommendations. Our experimental results on two real-world datasets reveal a substantial increase in aggregate diversity, a moderate increase in accuracy, reduced repetitiveness of recommendations, and demonstrate the importance of reinforcing diversity and novelty as complementary objectives.
  arXiv  Detail & Related papers  (2021-10-28T13:22:45Z)

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.