Is a Peeled Apple Still Red? Evaluating LLMs' Ability for Conceptual Combination with Property Type

TL;DR

This paper introduces a new dataset and evaluation framework for testing large language models' ability to perform conceptual combination involving property inheritance, emergence, and cancellation, revealing current limitations and proposing improvements.

Contribution

The study presents the CCPT dataset and tasks, along with a cognitive psychology-inspired method, to comprehensively evaluate and enhance LLMs' conceptual combination capabilities.

Findings

Automatic metric aligns with human judgments on property emergence and cancellation.

LLMs struggle to generate noun phrases with emergent properties.

Proposed method improves generative performance across tasks.

Abstract

Conceptual combination is a cognitive process that merges basic concepts, enabling the creation of complex expressions. During this process, the properties of combination (e.g., the whiteness of a peeled apple) can be inherited from basic concepts, newly emerge, or be canceled. However, previous studies have evaluated a limited set of properties and have not examined the generative process. To address this gap, we introduce the Conceptual Combination with Property Type dataset (CCPT), which consists of 12.3K annotated triplets of noun phrases, properties, and property types. Using CCPT, we establish three types of tasks to evaluate LLMs for conceptual combination thoroughly. Our key findings are threefold: (1) Our automatic metric grading property emergence and cancellation closely corresponds with human judgments. (2) LLMs, including OpenAI's o1, struggle to generate noun phrases which possess given emergent properties. (3) Our proposed method, inspired by cognitive psychology model that explains how relationships between concepts are formed, improves performances in all generative tasks. The dataset and experimental code are available at https://github.com/seokwon99/CCPT.git.