Y
Hacker News
new
|
ask
|
show
|
jobs
by
red75prime
133 days ago
It's relatively easy to analyze skill files. Shared chunks of neural networks (LLM adapters) can hide malicious behaviors better.