---
title: "基于构造驱动的注入：面向大语言模型的基于语言学的编辑代码混合指纹"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.25633v1"
permalink: "https://feed.astralor.com/items/itm_1794124684dbc77c0af5199b"
published: "2026-07-28T12:16:14.000Z"
score: 81
channel: "ai"
tags: ["代码混合","大语言模型","指纹注入","版权保护","跨语言对齐","零空间投影"]
---

# 基于构造驱动的注入：面向大语言模型的基于语言学的编辑代码混合指纹

## AI 点评

> 这篇论文在LLM版权保护领域迈出了关键一步，尤其解决了传统指纹的误触发和可过滤问题。对于AI模型开发者、平台方和法律合规人员来说，这是不可错过的最新方案。我留意到其代码混合指纹的设计巧妙，而联合优化思路更可能成为后续研究的基准。

## 速览

本文提出一种新型大语言模型版权保护指纹注入框架LCFEdit，通过联合优化指纹构造与注入两阶段，解决现有自然语言指纹易误触发、乱码指纹易被过滤的问题。该方法利用低资源语言构造代码混合指纹，结合跨语言对齐与零空间投影进行注入，在保持模型效用的同时实现鲁棒且不可感知的所有权验证。实验证明该方法在不可察觉性、可检测性和无害性上均优于基线。

## 关键点

- 提出统一框架LCFEdit，将指纹构造与注入联合优化，打破传统两阶段分离的局限。
- 利用低资源语言构造代码混合指纹，避免自然语言误触发和乱码被过滤的问题。
- 采用基于零空间投影和跨语言对齐的知识保留注入方法，确保模型性能不受损。
- 实验验证了该方法在不可察觉性、可检测性和无害性上的优势，为LLM版权保护提供新思路。

## 全文

Large language models (LLMs) are costly intellectual assets that remain exposed to unauthorized redistribution and commercial misuse. Injected fingerprints, i.e., trigger--target pairs embedded in model behavior, offer a practical, black-box-verifiable ownership signal, but existing methods decouple the two stages of the fingerprint life cycle: how a fingerprint is constructed and how it is injected. Existing fingerprinting frameworks suffer from two limitations. Natural-language fingerprints are prone to accidental activation, and garbled fingerprints are easily filtered by perplexity-based detection. Furthermore, decoupling construction from injection leaves the latter unaware of the trigger's linguistic structure, missing the opportunity for targeted optimization. We argue that fingerprint construction should drive injection, and present a unified fingerprinting framework that jointly optimizes both stages. First, LCF constructs code-mixing fingerprints by combining low-resource languages under a semantic-density substitution rule and grammar-biased mixing, yielding triggers whose perplexity sits far below garbled baselines while avoiding the accidental-activation failures of natural-language triggers. Second, LCFEdit injects each fingerprint with a null-space projection derived from high-resource multilingual representations that preserves knowledge, augmented by a cross-lingual alignment step that steers the weight update toward the fingerprint language's representation subspace. This construction-aware injection ensures that the update is linguistically informed and therefore more stable. Extensive evaluations on imperceptibility, detectability, and harmlessness demonstrate persistent ownership verification with negligible impact on utility.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_1794124684dbc77c0af5199b