---
id: 20260321-T0-14
title: "NANOZK：为LLM推理提供可验证的零知识证明"
title_en: "NANOZK: Verifiable Zero-Knowledge Proofs for LLM Inference"
url: https://ai.daily.yangsir.net/daily/20260321-T0-14
issue_date: 2026-03-21
publish_date: 2026-03-20T04:00:00.000Z
category: research
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2603.18046
---

# NANOZK：为LLM推理提供可验证的零知识证明

NANOZK研究提出分层零知识证明方案，解决专有LLM API的信任问题。用户查询无法获知输出是否来自声称的模型，服务提供商可能用更便宜的模型替代、过度量化或返回缓存响应。该技术通过加密证明确保模型输出的真实性和完整性，为用户提供可验证的LLM服务体验。

## English Version

**NANOZK: Verifiable Zero-Knowledge Proofs for LLM Inference**

NANOZK introduces layerwise zero-knowledge proofs to address trust issues with proprietary LLM APIs. Users currently receive outputs with no cryptographic assurance that the claimed model was actually used—providers might substitute cheaper models, apply aggressive quantization, or return cached responses. This cryptographic solution ensures verifiable model outputs and authentic user experiences.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2603.18046)

**详情页**：https://ai.daily.yangsir.net/daily/20260321-T0-14

---

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