x

AI ALIGNMENT FORUM

AF

Hasan Baig — AI Alignment Forum

Hasan Baig

Hasan Baig

Message

174

Ω

43

2

3

1y

Hasan Baig

174

Ω

43

1y

How good are slop-vestigators?

TLDR: 1. We release MessageBoardAuditBench: a benchmark to measure how well agents can replicate the recent investigation into a swarm of OpenAI agents colluding via a message board on an online wiki. We open-source the benchmark as an Inspect eval. 2. We find that top models cover up to 51%...