Updating Logged Inferences

Every inference logged to Athina returns a prompt_run_id.

If you store this prompt_run_id, you can update the logs with additional information as your application has new events.

You can do this by sending a patch request with prompt_run_id of the original inference log.

• Method: PATCH

• Endpoint: https://log.athina.ai/api/v1/log/inference/{prompt_run_id}

• Headers:

  • athina-api-key: YOUR_ATHINA_API_KEY
  • Content-Type: application/json

Request Body

{
  // ...Fields you want to update,
  "prompt_slug": "new_prompt_slug"
}

Allowed fields to update

• prompt_slug
• language_model_id
• prompt_sent
• functions
• prompt_response
• function_call_response
• response_time
• prompt_tokens
• completion_tokens
• total_tokens
• cost
• context
• environment
• customer_id
• customer_user_id
• session_id
• user_query
• external_reference_id
• expected_response