Non-exhaustive list of sources used in this course and evals tools that can help you.
Maud Nalpas
For more resources on testing and AI, we recommend the following resources.
- Learn Testing: Refresh your approach to testing.
- Learn AI: Design AI systems for your websites and web applications.
- Google DeepMind Evals: Multiple standardized benchmarking tools for different types on models
- Gemini Evaluations Playbook: Recipes for experimenting and evaluating generative AI models with Vertex AI.
- Responsible AI toolkit: Evaluate models and systems for safety.
- Evaluating your evals: A meta lesson on how to understand what evals to use, and what works effectively.
- Building better AI benchmarks: How many raters are enough? Understand an evaluation framework for ML models that optimizes the trade-off between the number of items and raters per item, to build reproducible AI benchmarks.
Course sources
We relied on several sources to write this series, including:
- AI Engineering: Building Applications with Foundation Models, Chip Huyen
- De-risking QA for LLM-powered applications by Michael Hablich, Chrome DevTools
- Using LLM-as-a-Judge For Evaluation: A Complete Guide by Hamel Husain
Eval tools
Examples of evals solutions and tools include:
- AlignEval
- Arize
- Braintrust
- Datadog
- DeepEval
- Gen AI evaluation service and API
- Inspect Evals
- JudgeLM
- LangSmith
- Evaluation harness
- OpenEvals
There are many more eval tools available. If you are using other tools, share them with us.