# 8K Long Context

L3-8B-Lunar-Stheno is a model merged from L3-8B-Lunaris-v1 and L3-8B-Stheno-v3.2, addressing issues of overly long responses and insufficient initiative while enhancing context awareness and text generation capabilities.

CogVLM2 is a multimodal dialogue model based on Meta-Llama-3-8B-Instruct, supporting both Chinese and English, with 8K context length and 1344*1344 resolution image processing capabilities.

Meta Llama 3 is a series of large language models developed by Meta, including 8B and 70B scales, focusing on dialogue optimization and safety.

TableLlama is an open-source general-purpose large model specifically designed for various table tasks, trained on the TableInstruct dataset and capable of handling contexts up to 8K in length.