In-Context vs Weight-Based Learning

Definition

Two forms of learning: in-context learning is sample efficient and fast but ephemeral and memory-heavy in a growing KV cache; weight-based learning is durable and compressive but sample-inefficient via gradient updates. Human learning compresses experience into weights.

Cast within

No topic-region cast yet — this appears once In-Context vs Weight-Based Learning's compiled claims are aligned into a topic region's argument tree.

Claims mentioning In-Context vs Weight-Based Learning (8)

Your notes

Loading notes...