OPENAI PLEDGES TO PUBLISH AI SAFETY TEST RESULTS MORE OFTEN

Summary from the TLDR InfoSec newsletter

OpenAI published a new Safety Evaluations Hub, a web page showing how the company's models score on various tests for harmful content generation, jailbreaks, and hallucinations. OpenAI has stated that it may add additional evaluations in the future.