arXiv cs.AI / cs.LG / cs.CL·17d agoForgetting Only What Matters: Layer-Selective Unlearning toward Robust LLMs#llm#machine-unlearning#memorizationAI research1
Hugging Face daily papers·20d agoContinual Learning Mechanisms Compose for Long-Horizon Memorization#catastrophic-forgetting#continual-learning#fine-tuning1
arXiv cs.AI / cs.LG / cs.CL·22d agoMolecular Déjà Vu: Digit-Level Retrieval of Published Values in Frontier Language Models#benchmark-contamination#llm-evaluation#memorizationAI research1
