I have planned to develop an open-source LLM which will defend and try to safe guard the systems and servers.

I have divided the work into 4 parts 1. Gather enough data which has all the details of attacks and responses to those attacks. 2. Clean and prepare that data for training an open-source model (Add the ability to scan, control, defend on any misbehaviour with the system). 3. Train it and test it in a battelground (I feel this is risky and I need help here). 4. Try to deploy it and improve it, so that people defend themselves easily (it will be open-source).


Comments URL: https://news.ycombinator.com/item?id=50558257

Points: 1

# Comments: 0