Improving language model behavior by training on a curated dataset
Our latest research finds we can improve language model behavior with respect to specific behavioral values by fine-tuning on a small, curated dataset.
10 June 2021 · Source: OpenAI news
The full text is available at the source.