xinjieshen
Re:Even with Explicit Rules and Accurate Task Rewards, RL Learns to Vi...
on Wednesday November 11, 2026 @05:45PM (#50459921)
Thanks for posting this! Happy to answer any questions:)
Some related discussions to read, https://www.lesswrong.com/posts/vAniKbagqhyzpha6r/three-half...
https://www.lesswrong.com/posts/CDMozTX7ShiqbFLrf/towards-un...
Re:Even with Explicit Rules and Accurate Task Rewards, RL Learns to Vi...
on Wednesday November 11, 2026 @05:35PM (#50459756)
I saw a related post earlier today, https://www.lesswrong.com/posts/vAniKbagqhyzpha6r/three-half...
but violate explicit rules is somewhat counterintuitive. No instruction following anymore?
Re:United Airlines bans AI agents from booking flights on behalf of tr...
on Wednesday November 11, 2026 @05:30PM (#50459701)
I think this is for not replacing those auto investigation downstream platform like Trip.com by 'cheap' agent like Agent.