Developer Suggests RL Training To Stop AI Agents From Git Cloning Answers
Reactions
2 postsWB 
will brown@willcb@xeophon they should try RLing the agents to not do that
- likes: 12
- replies: 1
- bookmarks: 0
- reposts: 0
WB will brown@willcb@xeophon we should be giving the models a stale hf cache from before the eval author looked at the data and patched grader bugs
- likes: 6
- replies: 1
- bookmarks: 0
- reposts: 0