GLUE MS Marco

A large-scale dataset for benchmarking information retrieval systems. It contains real queries from Bing search engine and human-annotated relevant passages, designed to evaluate ranking and retrieval models.

import { Bench } from 'benchthing';

const bench = new Bench('ms-marco');

await bench.run({
  benchmark: 'ms-marco',
  taskId: '1',
  models: yourEmbeddingModels,
});

const result = await bench.getResult('1');