|
|
|
|
|
by ErenalpCet
76 days ago
|
|
Really clever project. The self-referential loop is a great approach —
turning their scale against them. I've been thinking about the AI data
pipeline from the other side, building a memory filter for local LLMs
(MemoryGate), so seeing projects like this that target the scraping
stage is interesting. Have you considered adding noise variation to the
poison content so it's harder to fingerprint and filter out? |
|