Unverified50% confidenceSolutionExact time
A typical approach to query rewriting is to design a dedicated prompt template that passes the last N rounds of conversation history together with the current user input to an LLM, instructing it to output a semantically complete, independently retrievable query statement.
1
Sources
50%
Confidence
Long-term
Relevance
9/2/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTaking LLMs as an example, context includes the system prompt, user query, externally retrieved information, historical interaction records, and additional system outputs.74% similarUnverifiedExplicitly asking the model to search online for the latest information when posing a question can leverage instruction-following capabilities to force real-time retrieval.70% similarUnverifiedThe interaction of traditional LLM tools such as ChatGPT and Claude is essentially question-and-answer in nature, and generated code must be manually copied, pasted, and debugged.69% similarUnverifiedToolContext不仅能获取自定义元数据，还能访问当前的聊天记录State，包含用户提示词、大模型响应、工具调用参数等完整对话上下文信息68% similarUnverified语义缓存（Semantic Cache）通过计算新查询与历史查询的向量相似度，超过阈值时直接返回缓存答案，GPTCache等开源工具实现了这一机制，通常与TTL机制结合68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/843213API
curl https://kongchang.com/api/v1/knowledge/claims/843213MCP
get_claim(id=843213)