Improving instruction hierarchy in frontier LLMs

媒体报道 OpenAI 原文语言:英文

原文摘录

IH-Challenge trains models to prioritize trusted instructions, improving instruction hierarchy, safety steerability, and resistance to prompt injection attacks.

以下是来源提供的简短摘录,完整内容请查看原文。

参考来源

← 返回每日AI报告