tokyotech-llm/Llama-3-Swallow-8B-v0.1

Llama-3-Swallow-8B-v0.1 is an 8 billion parameter language model developed by tokyotech-llm, continually pre-trained from the Meta Llama 3 family with a significant addition of Japanese language data. This model features an 8192-token context length and is optimized for strong performance in both Japanese and English language tasks. It excels particularly in Japanese benchmarks, demonstrating enhanced cross-lingual adaptation capabilities.