Read an evaluation Call of a run

GET /v1/prompt-eval-suites/{prompt_eval_suite_id}/runs/{run_id}/calls/{call_id}

Answers one evaluation Call of the run (1.63.0): its scenario, arm, repetition and model family, how it ended, and its Turn metrics without text. transcript is null unless include=transcript adds the text of its Turns, which needs transcripts:read and makes the read an audited content read, as get_call_transcript is; this is what a calibration's labeller reads. An evaluation Call is a test Call, so a live API key is answered not_found, as for any test Call; a Call of another run is not_found; and a Call a worker before 1.63.0 reported answers no Turn and a null transcript.

Request

Set VOCAL_API to the base URL of your API and VOCAL_API_KEY to an API key with the prompts:read scope (Authentication and keys). Replace each placeholder, such as agt_... or <name>, with your own value.

curl "$VOCAL_API/v1/prompt-eval-suites/pes_.../runs/per_.../calls/call_..." \
  -H "Authorization: Bearer $VOCAL_API_KEY"

const response = await fetch(`${process.env.VOCAL_API}/v1/prompt-eval-suites/pes_.../runs/per_.../calls/call_...`, {
  headers: {
    Authorization: `Bearer ${process.env.VOCAL_API_KEY}`,
  },
});
console.log(response.status, await response.text());

import os

import requests

response = requests.get(
    f"{os.environ['VOCAL_API']}/v1/prompt-eval-suites/pes_.../runs/per_.../calls/call_...",
    headers={
        "Authorization": f"Bearer {os.environ['VOCAL_API_KEY']}",
    },
)
print(response.status_code, response.text)

Parameters

Responses

Fields of the 200 answer

The shape is PromptEvalCall.