1 link tagged with all of: optimization + agent-architecture
Click any tag below to further narrow down your results
Links
This article explains the concept of the "Context Tax" in large language models (LLMs) and offers strategies to minimize token usage and improve performance. It covers techniques like stable prefixes, append-only context, and using precise tools to enhance cache hits and reduce costs.