Powered by Vitest
Define and run evaluations the same way you run tests, with a familiar and intuitive interface
Define, run, and debug LLM evaluations with a familiar API
import { evaluate, scorers } from 'viteval';
evaluate('Color detection', {
data: async () => [
{ input: "What color is the sky?", expected: "Blue" },
{ input: "What color is grass?", expected: "Green" },
],
task: async (input) => {
const result = await generateText(input);
return result.text;
},
scorers: [scorers.levenshtein],
threshold: 0.8,
});
Viteval is free and open source,
made possible by wonderful sponsors.