Chanin, David; Hunter, Anthony; Camburu, Oana-Maria; (2024) Identifying Linear Relational Concepts in Large Language Models. In: Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. (pp. pp. 1524-1535). Association for Computational Linguistics: Mexico City, Mexico.

Preview

PDF 2024.naacl-long.85.pdf - Published Version Download (621kB) | Preview

Abstract

Transformer language models (LMs) have been shown to represent concepts as directions in the latent space of hidden activations. However, for any human-interpretable concept, how can we find its direction in the latent space? We present a technique called linear relational concepts (LRC) for finding concept directions corresponding to human-interpretable concepts by first modeling the relation between subject and object as a linear relational embedding (LRE) (Hernandez et al., 2023b). We find that inverting the LRE and using earlier object layers results in a powerful technique for finding concept directions that outperforms standard black-box probing classifiers. We evaluate LRCs on their performance as concept classifiers as well as their ability to causally change model output.

Download activity - last month

Export as CSVJSONXML

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item