Large Language Model GPT-4 Achieves Near-Human Performance in Public Turing Test

GPT-4, a large language model developed by OpenAI, achieved a 49.7% success rate in a public online Turing test, outperforming previous models but falling short of the human baseline of 66%.