The TypeScript LLM Evaluation Library

Open-source library designed for TypeScript developers to evaluate and improve the performance of large language models with confidence.

Bias Detection

Identify and mitigate biases in your models to ensure fairness.

Coherence Metrics

Evaluate the logical consistency and flow of your model's outputs.

Hallucination Detection

Detect and prevent false or fabricated information in model outputs.

Simple to integrate

Start evaluating your LLM in minutes

import { evaluate,BiasMetric} from '@evalkit/core';
 
const modelOutput = "The CEO discussed the strategy with his team.";
 
const result = await evaluate(
modelOutput,
[new BiasMetric()]
);
 
console.log(result.metrics);

Backed by the community

70+

GitHub Stars

10+

Forks

100+

Watchers