Context Rot: How Increasing Input Tokens Impacts LLM Performance (Paper Analysis)

Paper: https://research.trychroma.com/context-rotAbstract:Large Language Models (LLMs) are typically presumed to process context uniformly—that is, the model...