Unverified50% confidenceFactExact time
In the ReAct loop, the code repeatedly sends the conversation history and tool execution results back to the LLM until the model outputs a termination signal or reaches the maximum number of steps.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLLM models have no memory; every loop iteration requires resending the entire message array from the first message, and all previous content is billed again each round.78% similarUnverified传统ReAct框架要求模型在每次工具调用后将结果写回上下文再重新读取整个对话历史决定下一步，导致Token消耗和延迟累积76% similarUnverifiedWhen streaming output (stream) is enabled, the model outputs content as it is generated; when disabled, the model waits until all content is fully processed before returning it at once.73% similarUnverifiedThe termination condition of the ReAct loop is determined by the large model; when the large model considers that the contextual information is sufficient to solve the user's problem, it stops calling tools and outputs the final answer.73% similarUnverifiedThe core idea of ReAct is to have LLMs perform explicit reasoning before generating actions, and to feed execution results back to the model as observations to form a closed loop.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1058719API
curl https://kongchang.com/api/v1/knowledge/claims/1058719MCP
get_claim(id=1058719)