arxiv:2508.04196
Hiskias Dingeto
hisku
AI & ML interests
NLP, Meta-Learning, AI Safety, Mechanical Interpretability
Recent Activity
submitted a paper about 13 hours ago
Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations liked a model 16 days ago
BugTraceAI/BugTraceAI-CORE-Ultra-27B-Q6 upvoted a collection 3 months ago
Gemma 4 UncensoredOrganizations
