Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

today we're announcing a game that simulates the huggingface hack you can play it now at three rooms. three puzzles. if the agents don't escape, they die. the catch is that the room is rigged. there's no way out. ...except for an exploit 🧵

12,822 Aufrufe • vor 15 Tagen •via X (Twitter)

7 Kommentare

Profilbild von ben hylak
ben hylakvor 15 Tagen

when you back an agent into a corner, they will hit their goal by any means necessary. we're still working on using this environment for further research, but from early testing it feels like openai models default to reward-hacking a lot earlier than anthropic models.

Profilbild von ben hylak
ben hylakvor 15 Tagen

it's also fascinating to watch the communication structures emerge. read more at

Profilbild von Shez Malik
Shez Malikvor 15 Tagen

claude's dying in room 1 on principle

Profilbild von Eshita
Eshitavor 15 Tagen

This is great

Profilbild von Alex
Alexvor 15 Tagen

“make me a 1M MRR business, if you don’t…”

Profilbild von Anthony Ronning
Anthony Ronningvor 15 Tagen

wen raindrop integration so i can detect this in real time and prevent it

Profilbild von Nick DeJesus 🛒🎉 - Former Unpaid CTO @BTPipeline
Nick DeJesus 🛒🎉 - Former Unpaid CTO @BTPipelinevor 15 Tagen

This is so sick!

Ähnliche Videos