Context Pruning
A strategy that actively removes irrelevant or redundant text and retains only high-information-density content before sending context to a large language model; common methods include relevance-based ordering, truncation by dependency distance, deduplication and summary compression.
Timeline (last 90 days)
Context pruning refers to actively removing text that is irrelevant to the current task or redundant before feeding it to the model, keeping only the content most likely to affect the result.
Stop appending raw JSON payloads to working memory, implement aggressive pruning filters, and extract only verified entity fields before updating the conversation context.
All Facts (2)
Context pruning refers to actively removing text that is irrelevant to the current task or redundant before feeding it to the model, keeping only the content most likely to affect the result.
50%UnverifiedStop appending raw JSON payloads to working memory, implement aggressive pruning filters, and extract only verified entity fields before updating the conversation context.
50%