---
id: 20260612-T0-07
title: "PoQ-Judge：低成本评估去中心化LLM推理质量"
title_en: "PoQ-Judge evaluates decentralized LLM inference cheaply"
url: https://ai.daily.yangsir.net/daily/20260612-T0-07
issue_date: 2026-06-12
publish_date: 2026-06-11T04:00:00.000Z
category: research
source_name: "arXiv cs.CL (NLP)"
source_url: https://arxiv.org/abs/2606.11196
---

# PoQ-Judge：低成本评估去中心化LLM推理质量

arXiv论文提出PoQ-Judge框架，用于去中心化LLM推理网络的轻量级质量评估。该框架训练专用判别模型对查询-输出对打分，无需真实标签参考。实验显示该方法在保持评估准确性的同时显著降低计算成本，适合大规模PoQ验证场景。

## English Version

**PoQ-Judge evaluates decentralized LLM inference cheaply**

arXiv paper introduces PoQ-Judge, a lightweight evaluation framework for decentralized LLM inference. It trains judge models to score query-output pairs without ground-truth references, reducing computational costs while maintaining accuracy for large-scale PoQ validation.

---

**来源**：[arXiv cs.CL (NLP)](https://arxiv.org/abs/2606.11196)

**详情页**：https://ai.daily.yangsir.net/daily/20260612-T0-07

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*