Reasoning models struggle to control their chains of thought, and that’s good

Company news OpenAI Source language: English

OpenAI introduces CoT-Control and finds reasoning models struggle to control their chains of thought, reinforcing monitorability as an AI safety safeguard.

Excerpt

OpenAI introduces CoT-Control and finds reasoning models struggle to control their chains of thought, reinforcing monitorability as an AI safety safeguard.

Short excerpt from the source — open the original for the full text.

Sources

← Back to the daily brief