Search Papers | Poster Sessions | All Posters
Poster A91 in Poster Session A - Tuesday, August 6, 2024, 4:15 – 6:15 pm, Johnson Ice Rink
Extracting Object Feature Norms from Large Language Models
Stephen Mazurchuk1 (), Andrew Anderson1; 1Medical College of Wisconsin
Many cognitive neuroscience researchers collect semantic feature ratings for stimuli. While these ratings are often collected using online systems, it can be costly, and a considerable amount of work is sometimes required to prepare the query. Consequently, there is growing interest in augmenting this task by using the representations contained in large language models. We present and validate a method for extracting semantic ratings using a simple paradigm that allows researchers to easily estimate human norms using freely available open-source language models.
Keywords: LLM Semantic Feature Distributional