Skip to main content

🏢 Moonshot AI

MoBA: Mixture of Block Attention for Long-Context LLMs
·3939 words·19 mins· loading · loading
AI Generated 🤗 Daily Papers Natural Language Processing Large Language Models 🏢 Moonshot AI
MoBA: Mixture of Block Attention enables efficient long-context LLMs by dynamically selecting relevant blocks, improving performance without compromising efficiency.