AI-enhanced semantic feature norms for 786 concepts
This provides a more comprehensive dataset for cognitive science research, enabling better study of human conceptual knowledge, though it is incremental as it builds on existing norming methods with AI enhancement.
The paper tackled the problem of limited concept and feature coverage in semantic feature norms by augmenting human-generated norms with large language model responses, resulting in a dataset (NOVA) that shows higher feature density and overlap, and outperforms human-only norms and word-embedding models in predicting semantic similarity judgments.
Semantic feature norms have been foundational in the study of human conceptual knowledge, yet traditional methods face trade-offs between concept/feature coverage and verifiability of quality due to the labor-intensive nature of norming studies. Here, we introduce a novel approach that augments a dataset of human-generated feature norms with responses from large language models (LLMs) while verifying the quality of norms against reliable human judgments. We find that our AI-enhanced feature norm dataset, NOVA: Norms Optimized Via AI, shows much higher feature density and overlap among concepts while outperforming a comparable human-only norm dataset and word-embedding models in predicting people's semantic similarity judgments. Taken together, we demonstrate that human conceptual knowledge is richer than captured in previous norm datasets and show that, with proper validation, LLMs can serve as powerful tools for cognitive science research.