Saved in:
Bibliographic Details
Main Authors: Chen, Qian, Li, Dongyang, He, Xiaofeng
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.01644
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910727993819136
author Chen, Qian
Li, Dongyang
He, Xiaofeng
author_facet Chen, Qian
Li, Dongyang
He, Xiaofeng
contents Continuous prompts have become widely adopted for augmenting performance across a wide range of natural language tasks. However, the underlying mechanism of this enhancement remains obscure. Previous studies rely on individual words for interpreting continuous prompts, which lacks comprehensive semantic understanding. Drawing inspiration from Concept Bottleneck Models, we propose a framework for interpreting continuous prompts by decomposing them into human-readable concepts. Specifically, to ensure the feasibility of the decomposition, we demonstrate that a corresponding concept embedding matrix and a coefficient matrix can always be found to replace the prompt embedding matrix. Then, we employ GPT-4o to generate a concept pool and choose potential candidate concepts that are discriminative and representative using a novel submodular optimization algorithm. Experiments demonstrate that our framework can achieve similar results as the original P-tuning and word-based approaches using only a few concepts while providing more plausible results. Our code is available at https://github.com/qq31415926/CD.
format Preprint
id arxiv_https___arxiv_org_abs_2412_01644
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Concept Based Continuous Prompts for Interpretable Text Classification
Chen, Qian
Li, Dongyang
He, Xiaofeng
Computation and Language
Artificial Intelligence
Continuous prompts have become widely adopted for augmenting performance across a wide range of natural language tasks. However, the underlying mechanism of this enhancement remains obscure. Previous studies rely on individual words for interpreting continuous prompts, which lacks comprehensive semantic understanding. Drawing inspiration from Concept Bottleneck Models, we propose a framework for interpreting continuous prompts by decomposing them into human-readable concepts. Specifically, to ensure the feasibility of the decomposition, we demonstrate that a corresponding concept embedding matrix and a coefficient matrix can always be found to replace the prompt embedding matrix. Then, we employ GPT-4o to generate a concept pool and choose potential candidate concepts that are discriminative and representative using a novel submodular optimization algorithm. Experiments demonstrate that our framework can achieve similar results as the original P-tuning and word-based approaches using only a few concepts while providing more plausible results. Our code is available at https://github.com/qq31415926/CD.
title Concept Based Continuous Prompts for Interpretable Text Classification
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2412.01644