arXiv 2609.18310cs.CL

SEA-LION-v4.8：技术报告

SEA-LION-v4.8: A Technical Report

Adila Aulia, Ahmed Dabeer, Ahn Jeongmi, Antonyrex Sajeban, Chan Hok Teng Adwin, Cheng Zi Yi Nicholas, Choa Hsueh Mei Esther, Heng Jonathan, Jann Railey Estrada … 展开作者

Adila Aulia, Ahmed Dabeer, Ahn Jeongmi, Antonyrex Sajeban, Chan Hok Teng Adwin, Cheng Zi Yi Nicholas, Choa Hsueh Mei Esther, Heng Jonathan, Jann Railey Estrada Montalan, Lee Chwan Ren, Leong Wai Yi, Leong Wei Qi, Liew Rachel, Limkonchotiwat Peerat, Muhammad Ridzuan Bin Mokhtar, Nagarajan Karthik, Ng Boon Cheong Raymond, Ngee Chia Tai, Ngui Jian Gang, Nguyen Thanh Ngan, Ong Tat-Wee David, Pereira Mark, Phang Shi Wei Benjamin, Poon Joseph, Rengarajan Hamsawardhini, Susanto Yosephine, Sutaveephamochanon Anocha, Tan Choon Meng, Tan Chor Phin Evelyn, Tan Le Min Sheryl, Tan Siao Wei Jessica, Tan Yixian, Tasawong Panuthep, Tee Jun Yun, Teng Kok Wai Walter, Teo Eng Sipp Leslie, Tjhi William, Tuchinda Pume, Wu Donghang, Yong Xianbin, Zhang Zhou

首次发表 2026-09-21

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

浏览论文内容

中文总结 AI 辅助

本报告介绍基于NVIDIA Nemotron 3的SEA-LION-v4.8模型家族，通过多语言数据适配与后训练，在SEA-HELM上显著提升东南亚语言任务性能。

中文摘要 AI 辅助

我们推出了Nemotron-SEA-LION-v4.8，这是一个基于NVIDIA Nemotron 3构建的东南亚语言统一网络（SEA-LION）模型家族。该家族包括30B-A3B和120B-A12B两种模型，同时提供持续预训练的基础检查点和后训练变体。我们使用东南亚语言、推理、代码和多语言并行数据集对模型进行适配，随后通过监督微调和在线同策略蒸馏进行后训练。在SEA-HELM基准上，30B-A3B模型将整体东南亚语言得分从46.06提升至51.57，而120B-A12B模型则从49.30提升至63.44。在七种东南亚语言的指令遵循、自然语言推理和自然语言理解方面观察到了最显著的提升。

英文摘要

We introduce Nemotron-SEA-LION-v4.8, a family of Southeast Asian Languages In One Network (SEA-LION) models built upon NVIDIA Nemotron 3. The family includes 30B-A3B and 120B-A12B models, with both continued-pretrained base checkpoints and post-trained variants. We adapt the models using Southeast Asian, reasoning, code, and multilingual parallel data, followed by post-training with supervised fine-tuning and online on-policy distillation. On SEA-HELM, the 30B-A3B model improves the overall SEA score from 46.06 to 51.57, while the 120B-A12B model improves from 49.30 to 63.44. Across seven Southeast Asian languages, we observe broad capability gains with the 120B-A12B model showing broader and more consistent improvements across tasks.

发表机构

AI Singapore（新加坡全国人工智能计划）

机构由 AI 辅助整理，请以论文原文为准。

补充信息

备注: A technical report