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.

翻译：近年来，大语言模型（LLMs）的快速发展主要集中于英语领域，导致现有模型仅能使用英语进行响应。为了使这些模型适应其他语言，通常采用持续预训练（CP）方法，随后通过监督微调（SFT）以保持对话能力。然而，CP与SFT过程可能会削弱模型过滤有害内容的能力。本文提出指令持续预训练（InsCP），该方法在CP过程中引入指令标签，旨在使模型在习得新语言的同时不损失对话熟练度。实验表明，InsCP能够有效保留模型的对话能力及基于人类反馈的强化学习（RLHF）能力。在语言对齐性、可靠性及知识基准测试上的实证评估均证实了InsCP的有效性。值得注意的是，该方法仅需10亿标记的高质量指令遵循数据，显著降低了资源消耗。