Skip to content

Translation in progress Some entries still include text in the original language.

Failure-First·Indexed by this site · Originally published

Jailbreak-R1:用强化学习探索 LLM 的越狱能力

[Daily Paper] Jailbreak-R1: Exploring the Jailbreak Capabilities of LLMs via Reinforcement Learning

AI summary

Jailbreak-R1 提出用强化学习探索大语言模型的越狱能力,研究对象为 LLM 在越狱场景下的表现。

Read originalfailurefirst.org