Improving mathematical reasoning with process supervision

Company news OpenAI Source language: English

We’ve trained a model to achieve a new state-of-the-art in mathematical problem solving by rewarding each correct step of reasoning (“process supervision”) instead of simply rewarding the correct final answer (“outcome…

Excerpt

We’ve trained a model to achieve a new state-of-the-art in mathematical problem solving by rewarding each correct step of reasoning (“process supervision”) instead of simply rewarding the correct final answer (“outcome…

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

Sources

← Back to the daily brief