Open Access. Powered by Scholars. Published by Universities.®
Social and Behavioral Sciences Commons™
Open Access. Powered by Scholars. Published by Universities.®
Articles 1 - 2 of 2
Full-Text Articles in Social and Behavioral Sciences
Human Vs Machine: Hyper-Realistic Avatars And Their Efficacy As A Communication Channel, Jill S. Schiefelbein
Human Vs Machine: Hyper-Realistic Avatars And Their Efficacy As A Communication Channel, Jill S. Schiefelbein
USF Tampa Graduate Theses and Dissertations
Hyper-realistic avatars (HRAs), a form of synthetic media, are custom-created digital embodiments of a human, created by capturing and combining that person’s video and vocal likeness. This is the first known study of the efficacy of videos delivered by hyper-realistic avatars as a communication channel in comparison to videos delivered by their human counterparts. An experiment testing how information retention, engagement, and trust vary between viewers of videos delivered by a real human, videos delivered by the HRA representing that same human, and videos delivered by the HRA that discloses to viewers that it is a hyper-realistic avatar is presented. …
A Psychometric Analysis Of Natural Language Inference Using Transformer Language Models, Antonio Laverghetta Jr.
A Psychometric Analysis Of Natural Language Inference Using Transformer Language Models, Antonio Laverghetta Jr.
USF Tampa Graduate Theses and Dissertations
Large language models (LLMs) are poised to transform both academia and industry. But the excitement around these generative AIs has also been met with concern for the true extent of their capabilities. This dissertation helps to address these questions by examining the capabilities of LLMs using the tools of psychometrics. We focus on analyzing the capabilities of LLMs on the task of natural language inference (NLI), a foundational benchmark often used to evaluate new models. We demonstrate that LLMs can reliably predict the psychometric properties of NLI items were those items administered to humans. Through a series of experiments, we …