See full schedule

aiXamine (https://aixamine.qcri.org/) is a QCRI-developed platform for evaluating LLM safety and security. It includes 40+ tests to detect data and model vulnerabilities, such as backdoors, biases, adversarial attacks, and hallucinations. Users can submit models for evaluation, access leaderboards, compare models, and generate detailed reports. aiXamine helps ensure LLM reliability by identifying risks and assessing performance across key security and safety metrics.

aiXamine: LLM security and privacy, simplified