GEITje 7B Ultra:面向荷兰语的对话模型
GEITje 7B Ultra: A Conversational Model for Dutch
AI总结:
针对语言模型多聚焦英语、忽视其他语言的问题,本研究以源自Mistral 7B的荷兰语模型GEITje为基础,通过合成对话数据集监督微调与反馈数据集偏好对齐,扩展其能力并公开模型与数据集。
摘要:
语言模型发展迅速,但主要聚焦于英语,常忽视其他语言的大规模预训练。这种方法要求通过微调将强大的、以英语为中心的模型适配到其他语言环境。对于荷兰语,近期的一项尝试是“GEITje”,该模型最初源自基于英语的Mistral 7B。本研究在这一基础工作上,通过在新创建的高质量合成对话数据集上进行监督微调,以及在合成反馈数据集上进行额外的偏好对齐步骤,扩展了GEITje的能力。所开发的模型和创建的数据集均公开可用。
英文摘要:
Language models have rapidly evolved, predominantly focusing on English while often neglecting extensive pretraining in other languages. This approach has required initiatives to adapt powerful, English-centric models to other linguistic contexts through finetuning. For Dutch, such a recent endeavour is ``GEITje'' a model originally derived from the English-based Mistral 7B. Building on this fundamental work, the current research extends the capabilities of GEITje by supervised finetuning on newly created high-quality synthetic conversational datasets, along with an additional preference alignment procedure on a synthetic feedback dataset. Both the developed models and the created datasets are openly available.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。