OPENAI LAUNCHES SAFETY METRICS HUB AMID TRANSPARENCY CONCERNS

Summary from the TLDR AI newsletter

OpenAI has unveiled a Safety Evaluations Hub that publicly tracks model performance across hallucination rates, harmful content refusal, jailbreak resistance, and instruction adherence.