通过 Aident Loadout 使用 Metatextai Tools 的 Metatextai Evaluate。执行前查看输入、输出、设置、价格和风险。
Tool to evaluate LLM messages against policies/guardrails. Use after generating model output to get violation details or corrections.
messages必填arrayConversation messages to evaluate. Must include 'user' and 'assistant' messages; 'system' optional.
policiesarrayInline policies to evaluate; overrides console defaults when provided.
fail_fastbooleanIf true, stops evaluation on first violation.
policy_idsarrayList of policy IDs to override default console policies.
applicationstringApplication identifier; defaults to your configured application ID.
override_responsestringTop-level override response message on violation. Overrides correction.
correction_enabledbooleanIf true, returns corrected output when violations occur (unless override_response is set).
data必填Data from the action execution
errorstringError if any occurred during the execution of the action
successful必填booleanWhether or not the action execution was successful or not
API 密钥连接。风险等级 2/5。
免费操作
目前没有已发布技能明确引用此操作。