Which token marks the end of the tool results in function calling?

Study for the Hugging Face Agent Certification. Prepare with interactive quizzes and multiple-choice questions, complete with explanations and hints. Ace your exam!

Multiple Choice

Which token marks the end of the tool results in function calling?

Explanation:
In prompts that use tools, outputs from the tool are enclosed between markers. The token that marks the end of the tool results is the closing tag [/TOOL_RESULTS]. This signals that the tool has finished returning its data and the model can proceed with its own next steps or reasoning. The opening tag [TOOL_RESULTS] just starts that block, while other tokens like [TOOL_CALLS] or [AVAILABLE_TOOLS] serve different purposes (requesting tool usage or listing available tools) and do not indicate the end of the results. Having a clear end marker helps the system know where the tool output ends so parsing stays unambiguous.

In prompts that use tools, outputs from the tool are enclosed between markers. The token that marks the end of the tool results is the closing tag [/TOOL_RESULTS]. This signals that the tool has finished returning its data and the model can proceed with its own next steps or reasoning. The opening tag [TOOL_RESULTS] just starts that block, while other tokens like [TOOL_CALLS] or [AVAILABLE_TOOLS] serve different purposes (requesting tool usage or listing available tools) and do not indicate the end of the results. Having a clear end marker helps the system know where the tool output ends so parsing stays unambiguous.

Subscribe

Get the latest from Passetra

You can unsubscribe at any time. Read our privacy policy