Skip to main content

Interface: EvaluateConversationQualityParams

Properties​

deploymentName​

• deploymentName: string

The name of the OpenAI ChatGPT API deployment to use.

Example

"gpt-3.5-turbo"

Defined in​

packages/mongodb-chatbot-evaluation/src/evaluate/evaluateConversationQuality.ts:19

fewShotExamples​

• Optional fewShotExamples: ResponseQualityExample[]

Provide a few examples of conversation transcripts, expected outputs, and what the LLM output should be. This is extremely useful for helping the LLM understand its expected behavior for your use case. While not strictly necessary, it is highly recommended to include a few representative examples.

Here, the LLM utilizes a prompting technique called "few-shot prompting".

Defined in​

packages/mongodb-chatbot-evaluation/src/evaluate/evaluateConversationQuality.ts:31

openAiClient​

• openAiClient: OpenAIClient

Defined in​

packages/mongodb-chatbot-evaluation/src/evaluate/evaluateConversationQuality.ts:13