How far can you trust chain-of-thought prompting?
A high-tech news website that focuses on trends in machine learning, deep learning, neural networks and artificial intelligence, or AI, technologies and business, reports on a new research paper authored by Subbarao Kambhampati, a professor in the School of Computing and Augmented Intelligence, part of the Fulton Schools, along with graduate student Kaya Stechly and graduate research associate Karthik Valmeekam in the same school. In the paper, they examine the popular AI “chain-of-thought” prompting technique that improves performance but also has some limitations. The authors write about how to avoid the hazards presented by the technique.