22M parameter lightweight prompt safety classifier based on DeBERTa-xsmall, offering 75% reduced latency for detecting prompt injections and jailbreaks, primarily optimized for English.
Input: Text
Output: Text
Providers
groq
Credits
Context512
Max Output2
Input$0.010/1M
Output$0.010/1M
Cache Read—
Cache Write—
Quick Start
Use Meta Llama Prompt Guard 2 22M through Helicone's AI Gateway with automatic logging and monitoring.