Saved in:
Bibliographic Details
Main Authors: Chen, Kuang-Ming, Lee, Hung-yi
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2405.20175
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911894681419776
author Chen, Kuang-Ming
Lee, Hung-yi
author_facet Chen, Kuang-Ming
Lee, Hung-yi
contents The rapid development of large language models (LLMs) in recent years has largely focused on English, resulting in models that respond exclusively in English. To adapt these models to other languages, continual pre-training (CP) is often employed, followed by supervised fine-tuning (SFT) to maintain conversational abilities. However, CP and SFT can reduce a model's ability to filter harmful content. We propose Instruction Continual Pre-training (InsCP), which integrates instruction tags into the CP process to prevent loss of conversational proficiency while acquiring new languages. Our experiments demonstrate that InsCP retains conversational and Reinforcement Learning from Human Feedback (RLHF) abilities. Empirical evaluations on language alignment, reliability, and knowledge benchmarks confirm the efficacy of InsCP. Notably, this approach requires only 0.1 billion tokens of high-quality instruction-following data, thereby reducing resource consumption.
format Preprint
id arxiv_https___arxiv_org_abs_2405_20175
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle InstructionCP: A fast approach to transfer Large Language Models into target language
Chen, Kuang-Ming
Lee, Hung-yi
Computation and Language
Artificial Intelligence
The rapid development of large language models (LLMs) in recent years has largely focused on English, resulting in models that respond exclusively in English. To adapt these models to other languages, continual pre-training (CP) is often employed, followed by supervised fine-tuning (SFT) to maintain conversational abilities. However, CP and SFT can reduce a model's ability to filter harmful content. We propose Instruction Continual Pre-training (InsCP), which integrates instruction tags into the CP process to prevent loss of conversational proficiency while acquiring new languages. Our experiments demonstrate that InsCP retains conversational and Reinforcement Learning from Human Feedback (RLHF) abilities. Empirical evaluations on language alignment, reliability, and knowledge benchmarks confirm the efficacy of InsCP. Notably, this approach requires only 0.1 billion tokens of high-quality instruction-following data, thereby reducing resource consumption.
title InstructionCP: A fast approach to transfer Large Language Models into target language
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2405.20175