Return to Projects

Trusted evals request for proposals

ProjectTrust & Safety

We’re committed to providing the open-source AI community with a model evaluation platform that is repeatable, reproducible, diverse, and continuously improving as we discover new risks.

We are seeking new perspectives in the AI evaluation domain. If selected, your eval could be hosted on our Trusted Eval platform for others to leverage.

Our areas of interest include but are not limited to, cybersecurity threats, sensitive data detection, knowledge and factuality, multilingual evaluation, mediation, personal data memorization, data governance, areas related to CBRNE, weapons acquisition specifically, synthetic dataset integrity, demographic representation across different countries, distribution bias, political bias, violence and hate, terrorism and sanctioned individuals, and misinformation.

Request for proposals should include a link to a summary of your project in English, explaining the area of focus, dataset description, and any relevant prior work.

Submit your proposal