---
title: "GB/T-Bench：国家标准文档规则密集型审核的评测基准与增强方法"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.06312v1"
permalink: "https://feed.astralor.com/items/itm_0c1fb709004937241a2f75ff"
published: "2026-08-06T17:27:23.000Z"
score: 82
channel: "ai"
tags: ["arXiv","GB/T-Bench","GB/T-Reviewer","多智能体框架","评测基准"]
---

# GB/T-Bench：国家标准文档规则密集型审核的评测基准与增强方法

## AI 点评

> 这论文值得所有做文档智能与专业AI审核的人关注。它不只是又一个benchmark，而是把“规则密集型审核”这个过去严重依赖人力的场景，第一次拆成了可量化的任务。对标准化机构、法律科技和合规科技从业者来说，这套思路可以直接迁移到合同、法规、行业规范等场景。多智能体框架的收益也很有说服力——从0.328到0.509，说明结构化技能协调确实能撬动大模型的潜力。

## 速览

大型语言模型在复杂专业任务中应用日益广泛，但对规则密集型文档的审核能力缺乏系统评估。研究团队提出GB/T-Bench，这是首个面向国家标准文档结构化审核的评测基准，包含层级化审核分类法、25种可诊断错误类型和可控反例生成机制，基于488份文档生成7306个可追溯审核错误实例。同时提出多智能体框架GB/T-Reviewer，将审核知识转化为专业技能，协调全局检查、定向诊断、规则扫描和结果验证。实验显示最强模型仅达0.3280 CMCS，而专家为0.6640；GB/T-Reviewer将最佳成绩提升至0.5094，显著缩小人机差距，为高利害文档领域的可信AI奠定基础。

## 关键点

- GB/T-Bench是首个面向国家标准文档结构化审核的评测基准，定义25种可诊断错误类型。
- 通过确定性规则与大模型受限改写结合，生成7306个可追溯的审核错误实例。
- 实验显示最强通用大模型与专家差距显著，多智能体框架GB/T-Reviewer将最佳CMCS提升至0.5094。

## 全文

Large language models (LLMs) increasingly support complex professional tasks, yet their capabilities in rule-intensive document review remain insufficiently evaluated. National standard documents, such as China GB/T standards, offer a representative testbed: they are lengthy, highly structured, and governed by explicit rules for scope, terminology, normative wording, and cross-section consistency. Existing benchmarks focus on domain knowledge and question answering, largely overlooking intrinsic quality review for professional documents. Such reviews rely heavily on human experts, making them costly and difficult to scale. To bridge this gap, we introduce GB/T-Bench, the first benchmark for the structured review of national standard documents. Its GB/T Review Taxonomy is a hierarchical schema covering document structure, scope alignment, normative modality, terminology consistency, and normative references, with 25 diagnosable error types. A controllable counterexample generation mechanism combines deterministic rules and constrained LLM rewriting to process 488 documents into 7,306 traceable review error instances for evaluation. We also develop a diagnosis-oriented evaluation protocol requiring exact matches on error location, review dimension, and error type, plus document-level coverage metrics. We further propose GB/T-Reviewer, a multi-agent framework that converts review knowledge into specialized skills and coordinates global inspection, targeted diagnosis, rule scanning, and result verification. Experiments with 14 mainstream LLMs reveal a substantial human-LLM gap: the strongest model achieves only 0.3280 CMCS versus 0.6640 for experts. GB/T-Reviewer raises the best CMCS to 0.5094, showing the value of structured skill coordination for rule-intensive document review. This work paves the way for trustworthy AI in standardization and other high-stakes document domains.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_0c1fb709004937241a2f75ff