Research question
What did OpenAI learn from the staged release of GPT-2?
Key findings
- Documents misuse considerations, detection research and staged release rationale - Finds humans could not reliably distinguish GPT-2 text
Release Strategies and the Social Impacts of Language Models
OpenAI's report on GPT-2's staged release, the first structured release-strategy analysis for language models.
Contribute
Spot an error, or know a record we should add? Members can propose additions and submit corrections.