Gemma2 instruction-tuned architecture
The captured configuration identifies Gemma2ForCausalLM with model type gemma2 and Transformers support.
Open Source Model Profile · aisingapore
Gemma-SEA-LION-v3-9B-IT
Gemma-SEA-LION-v3-9B-IT is a 9.24B-parameter Gemma2 instruction-tuned model from AI Singapore. According to the model card, it targets Southeast Asian language use.
- Publisher
- aisingapore
- Task
- text-generation
- Model type
- gemma2
- License
- gemma
- Library
- transformers
- Publication status
- Approved for indexing
Model overview
Gemma-SEA-LION-v3-9B-IT is published by aisingapore as a Gemma2-based instruction-tuned text-generation model. The captured configuration identifies Gemma2ForCausalLM, and Safetensors metadata reports about 9.24B parameters under gemma. According to the model card, it continues Gemma-SEA-LION-v3-9B with English and ASEAN-language instruction tuning.
Recorded capabilities
Southeast Asia instruction tuning
According to the model card, SEA-LION models are pretrained and instruct-tuned for Southeast Asia, with this checkpoint tuned in English and ASEAN languages.
Thirteen documented languages
According to the model card, thirteen supported languages are listed across Burmese, Chinese, English, Filipino, Indonesian, regional languages, Tamil, Thai, and Vietnamese.
Documented evaluation scope
According to the model card, evaluation covers QA, sentiment, toxicity, translation, summarisation, reasoning, inference, and SEA-IFEval constraint following.
No system prompt or tool calling
According to the model card, this model was not trained to use a system prompt or tool calling.
Use cases in the source record
- Southeast Asian multilingual instruction-following work across the card's listed languages and task set.
- Translation, summarisation, QA, sentiment, and constraint-following experiments matching the card's documented evaluation scope.
Limitations and unknowns
- No structured evaluation scores were extracted from this record.
- No context-window value was extracted from this record.
- Provider state is historical snapshot data and should be refreshed before being presented as current.
- Training figures in the card describe the exact production run and may differ from aggregate totals or research-paper figures, per the publisher.
Source and provenance
Source: aisingapore/Gemma-SEA-LION-v3-9B-IT
Captured: Unknown. Processed: 2026-09-07T19:34:39.404753+00:00.
Current Version: 14.04.2025 Gemma-SEA-LION-v3-9B-IT SEA-LION is a collection of Large Language Models (LLMs) which have been pretrained and instruct-tuned for the Southeast Asia (SEA) region. SEA-LION stands for Southeast Asian Languages In One Network . Developed by: Products Pillar, AI Singapore Funded by: Singapore NRF Model type: Decoder Languages supported: Burmese, Chinese, English, Filipino, Indonesia, Javanese, Khmer, Lao, Malay, Sundanese, Tamil, Thai, Vietnamese License: Gemma Community License Model Details Model Description We performed instruction tuning in English and also in ASEAN languages such as Indonesian, Thai an…
F001F002F003F004F005F006F007F008F009F010F011F012F013F014F015F016F017F018F019F020