Improving instruction hierarchy in frontier LLMs

公司动态 OpenAI 原文语言:英文

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

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

← 返回每日简报