Deep Ignorance: Filtering Pretraining Data Builds Tamper-Resistant Safeguards into Open-Weight LLMs
Paper
•
2508.06601
•
Published
•
5
Kyle O'Brien PRO
Kyle1668
AI & ML interests
Interpretability, model editing, alignment
Recent Activity
updated
a model
about 17 hours ago
Unlearning/early-unlearning-weak-filter-ga-1-in-41-ga-lr-scale-0_001-gclip-0_5
published
a model
about 17 hours ago
Unlearning/early-unlearning-weak-filter-ga-1-in-41-ga-lr-scale-0_001-gclip-0_5
authored
a paper
6 days ago
Composable Interventions for Language Models
Organizations
