Llama Prompt Guard 2 86M

An 86M parameter BERT-based Meta classifier: catches prompt injections and jailbreak attempts, answering benign or malicious. Accepts a string of up to 512 tokens and handles English and non-English attacks; needs no specific prompt structure.

длинные запросы Meta советует разбивать на части, чтобы проверить целиком

Understands

text

This model cannot see images or video — you can only give it text. That matters if you planned to show it screenshots or documents with pictures.

Produces

text

Contents verified with the vendor: 2026-08-18

Included in subscriptions

No catalogue entry marks this model as part of a plan yet. It will appear here once the plan contents are confirmed with the vendor.