RSA-Control: A Pragmatics-Grounded Lightweight Controllable Text Generation Framework

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Yifan, Demberg, Vera
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912085211873280
author Wang, Yifan
Demberg, Vera
author_facet Wang, Yifan
Demberg, Vera
contents Despite significant advancements in natural language generation, controlling language models to produce texts with desired attributes remains a formidable challenge. In this work, we introduce RSA-Control, a training-free controllable text generation framework grounded in pragmatics. RSA-Control directs the generation process by recursively reasoning between imaginary speakers and listeners, enhancing the likelihood that target attributes are correctly interpreted by listeners amidst distractors. Additionally, we introduce a self-adjustable rationality parameter, which allows for automatic adjustment of control strength based on context. Our experiments, conducted with two task types and two types of language models, demonstrate that RSA-Control achieves strong attribute control while maintaining language fluency and content consistency. Our code is available at https://github.com/Ewanwong/RSA-Control.
format Preprint
id arxiv_https___arxiv_org_abs_2410_19109
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle RSA-Control: A Pragmatics-Grounded Lightweight Controllable Text Generation Framework
Wang, Yifan
Demberg, Vera
Artificial Intelligence
Computation and Language
Despite significant advancements in natural language generation, controlling language models to produce texts with desired attributes remains a formidable challenge. In this work, we introduce RSA-Control, a training-free controllable text generation framework grounded in pragmatics. RSA-Control directs the generation process by recursively reasoning between imaginary speakers and listeners, enhancing the likelihood that target attributes are correctly interpreted by listeners amidst distractors. Additionally, we introduce a self-adjustable rationality parameter, which allows for automatic adjustment of control strength based on context. Our experiments, conducted with two task types and two types of language models, demonstrate that RSA-Control achieves strong attribute control while maintaining language fluency and content consistency. Our code is available at https://github.com/Ewanwong/RSA-Control.
title RSA-Control: A Pragmatics-Grounded Lightweight Controllable Text Generation Framework
topic Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2410.19109