Tag
G-SHARE is a guideline-based structured reasoning framework for human-factor event diagnosis in nuclear power plants. It operationalizes a nine-step diagnostic guideline into a multi-stage pipeline with evidence extraction, stepwise reasoning, and consistency repair, outperforming one-shot LLM prompting and traditional baselines.
本文提出GradeSQL框架，使用结果奖励模型(ORM)进行Text-to-SQL的测试时验证，在BIRD和Spider基准上分别比基于执行的Best-of-N方法提升4.33%和2.10%。
Flow Reasoning Models (FRMs) introduce a training and test-time-scaling framework for discrete flow models on structured reasoning tasks. By using self-verification and self-conditioning, FRMs achieve nearly 100% solve rates on Sudoku and Zebra puzzles with far fewer passes than previous baselines.
This paper presents a five-arm ablation methodology for diagnosing which component of retrieval-warmed energy-based reasoning (RW-EBR) drives performance gains, applied to structured reasoning tasks like graph reachability and Sudoku. The method separates effects of class-prior bias, stochastic warm-starting, and graph-aligned value reuse.
ScaleToT proposes a method to generalize structured LLM reasoning for low-activity user modeling at billion scale, using tree-of-thought refinement and training a student model to reduce cost. An online A/B test in advertising deployment showed a 6.738% increase in LT30.
IBM released an open-source blog on Hugging Face detailing how to build robust enterprise agents with structured reasoning and tool use, going beyond basic LLMs and agents.
Proposes the Pseudocode-guided Structured Reasoning framework (PStar) that adaptively selects structured pseudocode reasoning paths to reduce hallucinations in Vision-Language Models, achieving state-of-the-art scores on POPE and MMStar benchmarks.