Материал на языке оригинала: English.
AI-written critiques help humans notice flaws
We trained “critique-writing” models to describe flaws in summaries. Human evaluators find flaws in summaries much more often when shown our model’s critiques. Larger models are better at self-critiquing, with scale improving critique-writing more than summary-writing. This shows promise for using AI systems to assist human supervision of AI systems on difficult tasks.
13 июня 2022 г. · Источник: OpenAI news
Полный текст доступен у источника.