Unverified50% confidenceFactExact time

该现象由开发者Simon Willison在其博客中重点解读

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

9/17/2026

First Seen

Valid until: 12/16/2026

Sources

OpenAI模型在压缩摘要中自我注入越狱指令

Related Entities

Simon Willisonperson

Related Claims

UnverifiedSimon Willison的博客是一个运行多年的Django项目71% similar UnverifiedSimon Willison的个人博客运行在自定义的Django应用上69% similar UnverifiedSimon Willison是「LLM」命令行工具的作者68% similar UnverifiedSimon Willison 在 Lobste.rs 社区讨论帖中分享了三篇对他职业思维方式产生深远影响的文章67% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/930340

API

curl https://kongchang.com/api/v1/knowledge/claims/930340

MCP

get_claim(id=930340)