Unverified90% confidenceFactTime unknown
LLM 库在 0.32 之前通过 conversation 对象管理多轮对话，但无法从外部注入已有的对话历史
1
Sources
90%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe mainstream LLM dialogue mechanism is essentially stateless, requiring the complete conversation history to be reprocessed with each API call.79% similarVerifiedMost current LLM applications start from scratch with each conversation, lacking persistent memory capabilities across sessions.71% similarUnverifiedTraditional large model conversations are stateless, and the context is lost once the conversation ends.70% similarUnverified传统对话历史模式的局限性包括：无法原生表达工具调用的中间状态、缺乏时序元数据、不支持并发事件流、在上下文窗口有限时难以进行选择性压缩和摘要65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/31286API
curl https://kongchang.com/api/v1/knowledge/claims/31286MCP
get_claim(id=31286)