Proactive LLM Security with resk-logits
RESK Security Links PyPI: pypi.org/project/resklogits GitHub: github.com/Resk-Security/resk-logits Site: resk.fr Jailbreaks bypass most LLM filters. That is why we built resk-logits. It is a GPU-accelerated Aho-Cor
RESK Security Links
- PyPI: pypi.org/project/resklogits
- GitHub: github.com/Resk-Security/resk-logits
- Site: resk.fr
Jailbreaks bypass most LLM filters. That is why we built resk-logits.
It is a GPU-accelerated Aho-Corasick processor that shadow-bans dangerous tokens at the logit level, BEFORE they are sampled.
Quick Start
from resklogits import LogitProcessor
# Configure processor
processor = LogitProcessor(patterns=["forbidden"])
# Filter logits during LLM inference
processed_logits = processor.process(logits)
Key Features
- GPU-accelerated pattern matching
- Under 1ms latency for 10,000+ patterns
- Real-time token shadow-ban
pip install resklogits to secure your LLM deployment today.
📰 Read the original article on Dev.to Security
Originally published by Dev.to Security. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.