註冊並分享邀請連結,可獲得影片播放與邀請獎勵。

Joachim Schaeffer
@JSchaeff3r
AI Safety Researcher working on AI Control and Security
加入 November 2021
203 正在關注    769 粉絲
Kimi K3.1 recalls instructions it was never given on Terminal Bench. They match the hidden tests. K3.1 while solving: "Instructions likely distilled pitfalls from previous agents." 🤯 When we edit a task to contradict them, K3.1 refuses the new instructions and fails. Other models follow the edit. Is it terminal(ly) benchmaxxed? 🧵
顯示更多
0
14
315
28
轉發到社區