Show HN: Prompt injection detector beats ProtectAI by 19% accuracy, 8.9x smaller(huggingface.co)
huggingface.co
Show HN: Prompt injection detector beats ProtectAI by 19% accuracy, 8.9x smaller
https://huggingface.co/hlyn/prompt-injection-judge-deberta-70m
3 コメント
Prompt injection will never be solved. Instruction following is the product.
Nothing in cybersecurity is truly solved—only made harder, Prompt injection is no different.
* 91.68% vs 72.28% accuracy
* 95.84% vs 65.33% precision — they block 1 in 3 legitimate users
* 83MB vs 739MB ONNX size
* 101ms vs 646ms on CPU