DynaGuard: A Dynamic Guardian Model With User-Defined Policies
This provides a more flexible tool for ensuring safety and ethical behavior in user-facing AI applications, though it is incremental over existing guardian models.
The authors tackled the problem of static guardian models for AI safety by introducing DynaGuard, a dynamic model that evaluates text based on user-defined policies, which surpasses static models in detection accuracy on traditional safety categories and is competitive with frontier reasoning models on free-form policy violations in a fraction of the time.
Guardian models play a crucial role in ensuring the safety and ethical behavior of user-facing AI applications by enforcing guardrails and detecting harmful content. While standard guardian models are limited to predefined, static harm categories, we introduce DynaGuard, a suite of dynamic guardian models offering novel flexibility by evaluating text based on user-defined policies, and DynaBench, a dataset for training and evaluating dynamic guardian models. Our models provide both rapid detection of policy violations and a chain-of-thought reasoning option that articulate and justify model outputs. Critically, DynaGuard not only surpasses static models in detection accuracy on traditional safety categories, but is competitive with frontier reasoning models on free-form policy violations, all in a fraction of the time. This makes DynaGuard an critical tool for language model guardrails.