Fine-tuning language models to find agreement among humans with diverse preferences

Open in new window