arXiv:2306.13649 — On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes

arXiv:2306.13649. v1 submitted 2023-06-23. Title on the /abs/ page confirmed as: "On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes". Venue/year for the corresponding paper record come from the Semantic Scholar graph API; see that record's notes where the arXiv comment field or OpenReview was used instead.