Model Releases
Interesting safety paper from Microsoft. They find that a weaker, unaligned model can split a harmful task into harmless-looking subquestion…
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 sess
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: https://academy.dair.ai/papers/divide-consult-conquer-capability-laundering-through-aligned-llms-2609.15383
Related
- Divide, Consult, Conquer: Capability Laundering Through Aligned LLMs
- Interesting new approach to enable memory in long-running agents. Weighted Memory Tree organizes execution into tasks, subtasks, and actions…
- Another big reason to use combination of frontier models. Chain-of-thought monitoring is treated as a reliable safety layer for agents. This…
Source: DAIR.AI (X) | 2026-09-16