A community-fine-tuned variant of Meta's Llama 3 8B, adapted by yentinglin with a focus on Traditional Chinese and Taiwanese cultural context. It handles instruction-following in both Chinese and English, with stronger grounding in Taiwan-specific language patterns than the base model. As an 8B parameter model, it operates within the constraints typical of that size class — capable and efficient, but not a match for larger models on complex reasoning tasks.