Topic: opt- testing phases

Sort by: Relevance | Date
  • OpenAI to Share More AI Safety Test Results Regularly
    May 14, 2025
    60%

    OpenAI to Share More AI Safety Test Results Regularly

    OpenAI launched a Safety Evaluations Hub to regularly publish detailed safety evaluation results for its AI models, covering harmful content, jailbreak attempts, and factual accuracy. The initiative aims to improve transparency and industry-wide understanding of AI safety...

    Read More »

Related Topics

ai model safety tests (1) harmful content metrics (1) jailbreak attempts (1) factual accuracy (1) criticism openai safety protocols (1) ai industry scrutiny (1) Transparency in AI (1) gpt-4o rollback (1)