DAIR.AI(@dair_ai):Interesting safety paper from Microsoft. They find that a weaker, unaligned model can split a harmful task into harmless-looking subquestions, ask an aligned frontier model each one in a separate session, and combine the answers locally. The authors call this capability laundering. Each request passes on its own, because no single answer from the frontier model is a harmful task. They tested GPT-5.5, Claude Opus 4.8 and Grok-4.3 as the consulted models. On CyBench, Gemma-4-31B recovered 8 of 14 tasks it failed alone when it consulted GPT-5.5. On a CBRN attack chain, consultation raised its mean rubric score from 62.3 to 83.1. Paper:

2026.09.16 10:20

Interesting safety paper from Microsoft. They find that a weaker, unaligned model can split a harmful task into harmless-looking subquestions, ask an aligned frontier model each one in a separate session, and combine the answers locally. The authors call this capability laundering. Each request passes on its own, because no single answer from the frontier model is a harmful task. They tested GPT-5.5, Claude Opus 4.8 and Grok-4.3 as the consulted models. On CyBench, Gemma-4-31B recovered 8 of 14 tasks it failed alone when it consulted GPT-5.5. On a CBRN attack chain, consultation raised its mean rubric score from 62.3 to 83.1. Paper:

显示更多