Saved in:
Bibliographic Details
Main Authors: Miyaoka, Yuya, Inoue, Masaki
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2408.15625
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914430270308352
author Miyaoka, Yuya
Inoue, Masaki
author_facet Miyaoka, Yuya
Inoue, Masaki
contents This paper proposes a control-based framework for aligning large language models (LLMs) by leveraging a control barrier function (CBF) to ensure user-desirable text generation. The presented framework applies the safety filter, designed based on the CBF, to the output generation of the baseline LLM, i.e., the sequence of the token, with the aim of intervening in the generated text. The overall text-generation system is implemented with Llama 3 and a RoBERTa model, and the source code is available at https://github.com/Mya-Mya/CBF-LLM. The experiment demonstrates its control ability and effectiveness in reducing the number of interventions needed for user-specified alignment tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2408_15625
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CBF-LLM: Safe Control for LLM Alignment
Miyaoka, Yuya
Inoue, Masaki
Systems and Control
Artificial Intelligence
Computation and Language
This paper proposes a control-based framework for aligning large language models (LLMs) by leveraging a control barrier function (CBF) to ensure user-desirable text generation. The presented framework applies the safety filter, designed based on the CBF, to the output generation of the baseline LLM, i.e., the sequence of the token, with the aim of intervening in the generated text. The overall text-generation system is implemented with Llama 3 and a RoBERTa model, and the source code is available at https://github.com/Mya-Mya/CBF-LLM. The experiment demonstrates its control ability and effectiveness in reducing the number of interventions needed for user-specified alignment tasks.
title CBF-LLM: Safe Control for LLM Alignment
topic Systems and Control
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2408.15625