Human-machine cooperation for semantic feature listing

TL;DR

This paper introduces a novel method that combines learned models and large language models to efficiently generate high-quality semantic feature norms, reducing the need for extensive human labor.

Contribution

It presents a new approach that integrates a learned model of human lexical-semantics with LLM-generated data for improved semantic feature listing.

Findings

Effective combination of models improves feature norm quality

Reduces human labor in generating semantic features

Achieves high-quality semantic feature lists

Abstract

Semantic feature norms, lists of features that concepts do and do not possess, have played a central role in characterizing human conceptual knowledge, but require extensive human labor. Large language models (LLMs) offer a novel avenue for the automatic generation of such feature lists, but are prone to significant error. Here, we present a new method for combining a learned model of human lexical-semantics from limited data with LLM-generated data to efficiently generate high-quality feature norms.