This essay explores whether contemporary Large Language Models (LLMs) can pass the Turing test, a benchmark proposed by Alan Turing to evaluate machine intelligence. The study involved evaluating four systems—GPT-4.5, LLaMa-3.1-405B, GPT-4o, and ELIZA—in randomized, controlled three-party Turing tests with two independent populations: UCSD undergraduate students and Prolific workers. Participants engaged in simultaneous conversations with a human and an AI system before judging which conversational partner they believed was human.

📁 Paper: https://arxiv.org/pdf/2503.23674


@scopeofai
@LLM_learning