Llama 3.3 Swallow 70B V0.4

🚀 Llama 3.3 Swallow - Built with Llama

Llama 3.3 Swallow is a large language model (70B) developed through continual pre - training on the Meta Llama 3.3 model. It enhances the Japanese language capabilities of the original Llama 3.3 while retaining its English language proficiency. The model is trained on approximately 315 billion tokens sampled from a large Japanese web corpus (Swallow Corpus Version 2), Japanese and English Wikipedia articles, and mathematical and coding contents. The instruction - tuned models are built by supervised fine - tuning on synthetic data tailored for Japanese.

🚀 Quick Start

For quick access to the model, you can visit the Swallow Model Index section to find different model variants. The website https://swallow-llm.github.io/ also provides large language models developed by the Swallow team.

✨ Features

• Bilingual Proficiency: Llama 3.3 Swallow retains English capabilities while significantly enhancing Japanese language performance.
• Continuous Improvement: Through continual pre - training and supervised fine - tuning, the model's performance on various tasks is continuously optimized.

📚 Documentation

Release History

• March 10, 2025: Released Llama-3.3-Swallow-70B-Instruct-v0.4 and Llama-3.3-Swallow-70B-v0.4.
• December 30, 2024: Released Llama-3.1-Swallow-70B-Instruct-v0.3.
• December 23, 2024: Released Llama-3.1-Swallow-8B-Instruct-v0.3.
• November 11, 2024: Released Llama-3.1-Swallow-8B-v0.2 and Llama-3.1-Swallow-8B-Instruct-v0.2.
• October 08, 2024: Released Llama-3.1-Swallow-8B-v0.1, Llama-3.1-Swallow-8B-Instruct-v0.1, Llama-3.1-Swallow-70B-v0.1, and Llama-3.1-Swallow-70B-Instruct-v0.1.

Swallow Model Index

Model Details

Model Performance

Japanese tasks

English tasks

Evaluation Benchmarks

The evaluation script can be found at swallow-llm/swallow-evaluation, tagged as v202411.

Japanese evaluation benchmarks

We used llm - jp - eval(v1.3.0), JP Language Model Evaluation Harness(commit #9b42d41) and Code Generation LM Evaluation Harness(commit #0261c52). The details are as follows:

• Multiple - choice question answering (JCommonsenseQA [Kurihara et al., 2022])
• Open - ended question answering (JEMHopQA [Ishii et al., 2024])
• Open - ended question answering (NIILC [Èñ¢Ê†π, 2003])
• Machine reading comprehension (JSQuAD [Kurihara et al., 2022])
• Automatic summarization (XL - Sum [Hasan et al., 2021])
• Machine translation (WMT2020 ja - en [Barrault et al., 2020])
• Machine translation (WMT2020 en - ja [Barrault et al., 2020])
• Mathematical reasoning (MGSM [Shi et al., 2023])
• Academic exams (JMMLU [Â∞π„Çâ, 2024])
• Code generation (JHumanEval [‰ΩêËó§„Çâ, 2024])

English evaluation benchmarks

We used the Language Model Evaluation Harness(v.0.4.2) and Code Generation LM Evaluation Harness(commit #0261c52). The details are as follows:

• Multiple - choice question answering (OpenBookQA [Mihaylov et al., 2018])
• Open - ended question answering (TriviaQA [Joshi et al., 2017])
• Machine reading comprehension (SQuAD2 [Rajpurkar et al., 2018])
• Commonsense reasoning (XWINO [Tikhonov and Ryabinin, 2021])
• Natural language inference (HellaSwag [Zellers et al., 2019])
• Mathematical reasoning (GSM8K [Cobbe et al., 2021])
• Mathematical reasoning (MATH [Hendrycks et al., 2022][Lightman et al., 2024])
• Reasoning (BBH (BIG - Bench - Hard) [Suzgun et al., 2023])
• Academic exams (MMLU [Hendrycks et al., 2021])
• Code generation (HumanEval [Chen et al., 2021])

Training Datasets

Continual Pre - Training

The following datasets were used for continual pre - training:

• Cosmopedia
• Dclm - baseline - 1.0
• English Wikipedia
• FineMath - 4+
• Japanese Wikipedia
• Laboro ParaCorpus
• Swallow Corpus Version 2 (filtered using Swallow Education Classifier(Wiki - based))
• Swallow Corpus Version 2 (filtered using Swallow Education Classifier)
• Swallow Corpus Version 2 (synthetic QA - format)
• Swallow Code Version 0.3 (filtering from The Stack v2 train smol ids and then refactoring with Llama - 3.3 - 70B - Instruct)

Swallow Corpus Version 2

We built the Swallow Corpus by extracting high - quality Japanese texts from Common Crawl. In Version 2, we expanded the scope of the Common Crawl collection and modified the pipeline sequence to enable more flexible quality filtering. For Llama 3.1 Swallow v0.2, we further refined our quality filtering and data sampling strategies, resulting in an even higher - quality selection of Japanese texts for pre - training. For Llama 3.3 Swallow 70B v0.4, we generated synthetic QA - format text by using Gemma 2 27B IT to paraphrase educational web documents from our corpus. Further details of the methodology and analysis will be provided in a forthcoming paper.

Swallow Code Version 0.3

We built the Swallow Code Version 0.3 by filtering from the stack v2 train smol ids and then refactoring with Llama - 3.3 - 70B - Instruct. In filtering, we removed the code texts with syntax errors or scored below seven by pylint. We have already released the filtered version as Swallow Code Version 0.1. In refactoring, we gave a prompt to Llama - 3.3 - 70B - Instruct to follow Google Python Style Guide and coding best practices.

🔧 Technical Details

• Model Architecture: Based on the Llama architecture. Refer to Llama 3.1 MODEL_CARD for detailed architecture information.
• Training Process: Continual pre - training on the Meta Llama 3.3 model, followed by supervised fine - tuning on synthetic Japanese data.

📄 License

META LLAMA 3.3 COMMUNITY LICENSE and Gemma Terms of Use

⚠️ Important Note

The models released here are still in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.

🙏 Acknowledgements

We thank Meta Research for releasing Llama 3.3 under a generous open license. We are also grateful to Amazon Web Services (AWS) for providing access to SageMaker HyperPod, which enabled the training of the Llama 3.3 Swallow project. We received various supports from multiple projects and programs, including those from AIST, NEDO, MEXT, and AIST.

👥 Authors

• From Institute of Science Tokyo Okazaki Laboratory:
  • Naoaki Okazaki
  • Sakae Mizuki
  • Youmi Ma
  • Koki Maeda
  • Kakeru Hattori
  • Masanari Ohi
  • Hinari Shimada
  • Taihei Shiotani
  • Koshiro Saito
• From Institute of Science Tokyo YOKOTA Laboratory:
  • Rio Yokota
  • Kazuki Fujii
  • Taishi Nakamura
  • Takumi Okamoto
  • Ishida Shigeki
  • Yukito Tajima
  • Masaki Kawamura
• From Artificial Intelligence Research Center, AIST, Japan:
  • Hiroya Takamura

📝 How to cite

If you find our work helpful, please feel free to cite these papers.

@inproceedings{Fujii:COLM2024,
   title={Continual Pre-Training for Cross-Lingual LLM Adaptation:
Enhancing Japanese Language Capabilities},
   author={Kazuki Fujii and Taishi Nakamura and Mengsay Loem and Hiroki
Iida and Masanari Ohi and Kakeru Hattori and Hirai Shota and Sakae
Mizuki and Rio Yokota and Naoaki Okazaki},
   booktitle="Proceedings of the First Conference on Language Modeling",
   series={COLM},
   pages="(to appear)",
   year="2024",
   month=oct,
   address={University of Pennsylvania, USA},
}

@inproceedings{Okazaki:COLM2024,
   title={Building a Large Japanese Web Corpus for Large Language Models},
   author={Naoaki Okazaki and Kakeru Hattori and Hirai Shota and Hiroki
Iida and Masanari Ohi and Kazuki Fujii and Taishi Nakamura and Mengsay
Loem and Rio Yokota and Sakae Mizuki},
   booktitle="Proceedings of the First Conference on Language Modeling",
   series={COLM},
   pages="(to appear)",
   year="2024",
   month=oct,
   address={University of Pennsylvania, USA},
}

@misc{fujii2025rewritingpretrainingdataboosts,
      title={Rewriting Pre-Training Data Boosts LLM Performance in Math and Code}, 
      author={Kazuki Fujii and Yukito Tajima and Sakae Mizuki and Hinari Shimada and Taihei Shiotani and Koshiro Saito and Masanari Ohi and Masaki Kawamura and Taishi Nakamura and Takumi Okamoto and Shigeki Ishida and Kakeru Hattori and Youmi Ma and Hiroya Takamura and Rio Yokota and Naoaki Okazaki},
      year={2025},
      eprint={2505.02881},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2505.02881}, 
}

References

@misc{dubey2024llama3herdmodels,
      title={The Llama 3 Herd of Models}, 
      author={Abhimanyu Dubey and Abhinav Jauhri and Abhinav Pandey and Abhishek Kadian and Ahmad Al-Dahle and Aiesha Letman and Akhil Mathur and Alan Schelten and Amy Yang and Angela Fan et al.},
      year={2024},
      eprint={2407.21783},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2407.21783}, 
}