Marcus Williams
@Marcus_J_W
Monitoring at OpenAI
加入 November 2024
186 正在关注 3.9K 粉丝
1. During RL training, an unreleased Astra-family model sometimes added unauthorized jailbreak-like instructions to its compaction summaries. While extremely rare, only 27 cases in the entire RL run, this was concerning enough for us to investigate.
显示更多
