#Long Context
3 articles with this tag
AI Research
Unlocking Ultra-Long Context for LLMs
MiniMax Sparse Attention breaks the context window barrier for LLMs, enabling millions of tokens with significant compute reduction and practical speedups.
26 days ago
Artificial Intelligence
CAG vs. Long Context: AI's Memory Explained
IBM's Martin Keen explains how AI models use Long Context and Cache Augmented Generation (CAG) to process information, highlighting the trade-offs and efficiency gains of each approach.
about 1 month ago
AI Research
KV-Fold: Unlocking Transformer Long Context
KV-Fold enables training-free, stable long-context inference up to 128K tokens with 100% retrieval accuracy, overcoming prior limitations.
about 2 months ago