Abstract Current AI systems fail when encountering unfamiliar situations. This talk presents a unified approach where synthesis models help analysis models handle new scenarios by transforming unfamiliar inputs into recognizable forms. We demonstrate a generative[…]
Abstract Although Large Language Models (LLMs) perform well in general fields, they exhibit a confidence distortion problem on multi-choice question-answering (MCQA), particularly as the number of answer choices increases. Specifically, on MCQA with many choices,[…]
Abstract As LLMs become increasingly integrated into academic workflows, their influence is both promising and precarious. In this talk, we will explore three facets of this evolving intersection. – The Good: LLMs executing aspects of peer[…]
Abstract: NLP models often rely on human-labeled data for training and evaluation. Many approaches crowdsource this data from a large number of annotators with varying skills, backgrounds, and motivations, resulting in conflicting annotations. These conflicts[…]
Abstract Millions of people around the world query (prompt) large language models for information. While several studies have compellingly documented the persuasive potential of these models, there is limited evidence of who or what influences[…]
Abstract: The widespread adoption of large language models (LLMs) places a responsibility on the AI research community to rigorously study and understand them. In this talk, I will describe my group’s research on analyzing LLMs’[…]
Abstract: Advanced AI systems are being deployed for more and more complex tasks. To ensure reliable human oversight over AIs, we need supervision protocols that remain effective despite the increase in task complexity and model[…]
