Human-in-the-Loop
A mechanism in Agent applications that pauses execution when a process reaches a point requiring external input, waiting for human intervention before continuing.
Timeline (last 90 days)
In the HR recruitment Agent, after the interview questions are generated and sent to the applicant, the process pauses to wait for the answer; this mechanism is called human intervention or breakpoint resumption, and the process continues only after the answer is obtained.
The evaluation pyramid strategy suggests: what can be resolved with rules and automated assertions should not all be handed to an LLM; semantic understanding and open-ended tasks should be judged by the model; and a small number of high-risk results should be handled with human intervention.
A safe agent should recognize when it must stop rather than continue chasing the original goal; in medical scenarios, an agent that proactively halts and requests human intervention may be more valuable than an agent that forcibly completes the task.
When a human steps in and resolves an error by other means, a negative signal should be sent, and the human fix should be written as a new note into the memory the bot recalls.
All Facts (4)
In the HR recruitment Agent, after the interview questions are generated and sent to the applicant, the process pauses to wait for the answer; this mechanism is called human intervention or breakpoint resumption, and the process continues only after the answer is obtained.
50%UnverifiedThe evaluation pyramid strategy suggests: what can be resolved with rules and automated assertions should not all be handed to an LLM; semantic understanding and open-ended tasks should be judged by the model; and a small number of high-risk results should be handled with human intervention.
50%UnverifiedA safe agent should recognize when it must stop rather than continue chasing the original goal; in medical scenarios, an agent that proactively halts and requests human intervention may be more valuable than an agent that forcibly completes the task.
50%UnverifiedWhen a human steps in and resolves an error by other means, a negative signal should be sent, and the human fix should be written as a new note into the memory the bot recalls.
50%