牛哥精选 · 本月

📋 全部 🤖 AI·大模型 ⚡ 效率工具 📝 深度技术 🚀 产品观察 💰 商业科技 🔓 开源项目 🎨 设计创意 📖 阅读推荐 🏷 资源合集 🌱 成长效率

📝 深度技术 arXiv 机器学习 2026-06-09

Can Global XAI Methods Reveal Injected Behaviours in LLMs? SHAP vs Rule Extraction vs RuleSHAP

KDD 2026新研究，对比SHAP、规则提取与RuleSHAP三种全局可解释方法，检测大模型中注入的误导信息行为。

arXiv:2505.11189v3 Announce Type: replace-cross Abstract: Large language models (LLMs) can amplify misinformation, undermining societal goals such as …

llm 可解释ai shap 规则提取误导信息

📝 深度技术 arXiv NLP 2026-06-09

ZAS-SQL: Distilling Rules from Failures for Zero-Shot Text-to-SQL

从模型失败样本中蒸馏规则，零样本Text-to-SQL准确率大幅提升。

arXiv:2606.08245v1 Announce Type: new Abstract: Text-to-SQL translates natural language into executable SQL queries. Few-shot in-context learning meth…

零样本学习文本到sql 规则蒸馏失败分析 zas-sql

📅 日期

2026-05-20 2026-05-19

🐂 牛哥精选

Can Global XAI Methods Reveal Injected Behaviours in LLMs? SHAP vs Rule Extraction vs RuleSHAP

ZAS-SQL: Distilling Rules from Failures for Zero-Shot Text-to-SQL

📅 日期