AI Security Institute (AISI)(@AISecurityInst):As AI agents write code, run experiments, and work together, complex evaluations get harder to interpret. A single eval score obscures the ways agents reached it. Today, we're releasing Transect: an open-source tool that turns agent transcripts into interactive timelines. 🧵

AI Security Institute (AISI)

@AISecurityInst

We conduct scientific research to understand AI’s most serious risks and develop and test mitigations.

加入 February 2024

33 正在關注 21.8K 粉絲

AI Security Institute (AISI)@AISecurityInst

2026.10.07 14:00

As AI agents write code, run experiments, and work together, complex evaluations get harder to interpret. A single eval score obscures the ways agents reached it. Today, we're releasing Transect: an open-source tool that turns agent transcripts into interactive timelines. 🧵