Tag
AdaPRA introduces a persistent coding agent harness with adaptive rubric-guided process supervision that accumulates knowledge across attempts, boosting an 80B Qwen model's leaderboard percentile by 24.5 points on DSPredict-Easy and 14.8 on DSPredict-Hard while reducing process errors in data science agents.
Fudan University 等机构提出 AiO-Mem，将记忆智能体的搜索、写入和停止操作统一到共享证据状态与奖励空间中，通过过程奖励模型与 PA-DAPO 优化记忆策略，在长上下文记忆基准上持续提升答案准确性。
LeanPRM提出利用Lean证明助手自动为自然语言数学证明生成步骤级正确性标签，无需人工标注即可训练过程奖励模型（PRM）。Qwen3-4B验证器在该数据上训练后，步级标签准确率从48.2提升至74.7，并在IMO级问题的最佳-of-N选择和多个外部验证基准上超越PRM800K训练的验证器。
The paper shows that deterministic PRM guidance in discrete diffusion language models underperforms simpler baselines when compute is matched, identifying two key failures: weak pruning signals and poor final judgment, and releases a corpus and toolkit for reproducible comparisons.
PRO-Step introduces a step-level process reward optimization method for retrieval-augmented generation to improve multi-hop reasoning by evaluating logical validity and evidential grounding at each step.
The paper proposes a neuro-symbolic framework that decouples reasoning into symbolic validity and semantic groundedness, using a verifier and a trained PRM to improve reliability in scientific reasoning tasks for LLMs.
KV-PRM introduces a process reward model that leverages KV-cache transfer to avoid re-encoding, achieving up to 5000x FLOP reduction while maintaining or improving performance on reasoning benchmarks.
Introduces MARD, a 7B-parameter model for mechanism-level drug-drug interaction prediction using mirror-augmented reasoning distillation, achieving state-of-the-art accuracy at ~1% of frontier API cost and demonstrating genuine pharmacological reasoning over memorization.
This paper introduces Multimodal Multi-Dimensional Scalarization Process Reward Modeling (MMS-PRM), which enforces the worst dimension's robustness in multimodal reasoning to prevent failures like visual hallucinations from being masked by strong text logic.
SCI-PRM introduces a tool-aware Process Reward Model for scientific reasoning, trained on the SCIPRM70K dataset featuring 'Chain-of-Tool' trajectories that interleave reasoning with scientific tool execution. It enables effective test-time scaling and serves as a dense reward signal in reinforcement learning, outperforming proprietary models like GPT-5-Mini on tool-calling steps across scientific benchmarks.
This paper introduces a framework for time series forecasting that uses importance-aware news compression and process reward model-guided retrieval to incorporate long news articles within fixed context limits, improving prediction accuracy across finance, energy, traffic, and Bitcoin benchmarks.
This paper proposes MERIT, a dynamic multi-horizon memory retrieval framework for interactive text-to-SQL agents that uses episode-level and turn-level memory with learned retrieval policies optimized via reinforcement learning and a process reward model for dense rewards. Experiments on BIRD-Interact and Spider2-Snow show that MERIT outperforms static and single-horizon dynamic baselines in success rate while requiring fewer interaction turns.
BetaPRM is a process reward model that predicts both a step-level success probability and the reliability of that prediction using a Beta belief from Monte Carlo continuations, enabling adaptive computation allocation that reduces token usage by up to 33.57% while improving accuracy.