agent_eval
autogen.agentchat.contrib.agent_eval.agent_eval.generate_criteria
generate_criteria
Creates a list of criteria for evaluating the utility of a given task.
Parameters:Name | Description |
---|---|
llm_config | llm inference configuration. Type: dict | Literal[False] | None Default: None |
task | The task to evaluate. Type: autogen.agentchat.contrib.agent_eval.task.Task Default: None |
additional_instructions | Additional instructions for the criteria agent. Type: str Default: ” |
max_round=2 | |
use_subcritic | Whether to use the subcritic agent to generate subcriteria. Type: bool Default: False |