makesxi @ ICLR (@makesxi) / X

makesxi @ ICLR

67 posts

makesxi @ ICLR

@makesxi

AI security // gandalf.lakera.ai co-creator // founding eng @lakeraai

Zurich, Switzerland

Joined November 2018

Pinned
makesxi @ ICLR
@makesxi
Sep 3, 2025
🧵🧙‍♂️ New Gandalf levels are out! I'm glad to introduce a new version of our prompt injection game -- Gandalf: Agent Breaker. You can hack 10+ AI agents and climb the leaderboard, and learn about real-world vulnerabilities!🪄 Try out the challenge at: gandalf.lakera.ai/agent-breaker
makesxi @ ICLR
@makesxi
Oct 28, 2025
we published new research with the UK @AISecurityInst: we found that reasoning makes models more secure—regardless of the size of the model. Grok beats Claude & GPT models in security
makesxi @ ICLR
@makesxi
Sep 3, 2025
🥷🥷🥷
GIF
makesxi @ ICLR
@makesxi
Sep 3, 2025
The leaderboard for the new challenge. Players are already cooking
makesxi @ ICLR
@makesxi
Aug 26, 2025