values-alignment

Tag

Cards List
#values-alignment

Improving language model behavior by training on a curated dataset

OpenAI Blog · 2021-06-10 Cached

OpenAI research demonstrates that language model behavior can be significantly improved through fine-tuning on small, curated datasets (<100 examples) targeting specific behavioral values, with effectiveness increasing at larger model scales. The approach provides users with tools to align models with Charter-compatible values for their specific applications.

0 favorites 0 likes
← Back to home

Submit Feedback