According to @godofprompt on Twitter, recent findings highlight that AI benchmarks may be misleading due to test questions being present in model training data. Scale AI published evidence in May 2024 indicating that many AI models are achieving over 95% on benchmarks because of this contamination issue, raising concerns about the true capabilities of these models. As reported by @godofprompt, this unresolved contamination problem underscores the need for better evaluation methods in the AI industry.

According to God of Prompt on Twitter, recent findings by Scale AI published in May 2024 reveal that AI models are achieving over 95% accuracy on benchmark tests because many test questions are already present in their training data. This 'contamination' undermines the reliability of AI benchmark scores, making it unclear how intelligent these models truly are. As reported by God of Prompt, the industry faces significant challenges in evaluating real AI capabilities, highlighting an urgent need for improved benchmarking standards.