On the Tip of the Tongue: Analyzing Conceptual Representation in Large Language Models with Reverse-Dictionary Probe

2024/02/22 by Xu, Ningyu, Zhang, Qi, Zhang, Menghan +2 · 1 citation
#Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2402.14404

Abstract

Probing and enhancing large language models' reasoning capacity remains a crucial open question. Here we re-purpose the reverse dictionary task as a case study to probe LLMs' capacity for conceptual inference. We use in-context learning to guide the models to generate the term for an object concept implied in a linguistic description. Models robustly achieve high accuracy in this task, and their representation space encodes information about object categories and fine-grained features. Further experiments suggest that the conceptual inference ability as probed by the reverse-dictionary task predicts model's general reasoning performance across multiple benchmarks, despite similar syntactic generalization behaviors across models. Explorative analyses suggest that prompting LLMs with description⇒word examples may induce generalization beyond surface-level differences in task construals and facilitate models on broader commonsense reasoning problems.

Cited by

• Understanding Subword Compositionality of Large Language Models

Related

• Unpacking Large Language Models with Conceptual Consistency
• Navigating Semantic Relations: Challenges for Language Models in Abstract Common-Sense Reasoning
• Decoding Knowledge in Large Language Models: A Framework for Categorization and Comprehension
• COPEN: Probing Conceptual Knowledge in Pre-trained Language Models
• Delving into the Reversal Curse: How Far Can Large Language Models Generalize?