Ranking entities along conceptual space dimensions with LLMs: An analysis of fine-tuning strategies

Kumar, Nitesh, Chatterjee, Usashi and Schockaert, Steven ORCID: https://orcid.org/0000-0002-9256-2881 2024. Ranking entities along conceptual space dimensions with LLMs: An analysis of fine-tuning strategies. Presented at: The 62nd Annual Meeting of the Association for Computational Linguistics, Bangkok, Thailand, 11-16 August 2024. Published in: Ku, Lun-Wei, Martins, Andre and Srikumar, Vivek eds. Findings of the Association for Computational Linguistics ACL 2024. Association for Computational Linguistics, pp. 7974-7989.

Preview

PDF - Published Version Download (449kB) | Preview

Abstract

Conceptual spaces represent entities in terms of their primitive semantic features. Such representations are highly valuable but they are notoriously difficult to learn, especially when it comes to modelling perceptual and subjective features. Distilling conceptual spaces from Large Language Models (LLMs) has recently emerged as a promising strategy, but existing work has been limited to probing pre-trained LLMs using relatively simple zero-shot strategies. We focus in particular on the task of ranking entities according to a given conceptual space dimension. Unfortunately, we cannot directly fine-tune LLMs on this task, because ground truth rankings for conceptual space dimensions are rare. We therefore use more readily available features as training data and analyse whether the ranking capabilities of the resulting models transfer to perceptual and subjective features. We find that this is indeed the case, to some extent, but having at least some perceptual and subjective features in the training data seems essential for achieving the best results.

Item Type:	Conference or Workshop Item (Paper)
Date Type:	Publication
Status:	Published
Schools:	Computer Science & Informatics
Publisher:	Association for Computational Linguistics
Date of First Compliant Deposit:	30 July 2024
Date of Acceptance:	16 May 2024
Last Modified:	19 Aug 2024 09:47
URI:	https://orca.cardiff.ac.uk/id/eprint/170092

Actions (repository staff only)

Edit Item