M-Loss: Quantifying Model Merging Compatibility with Limited Unlabeled Data

Documenti analoghi

• MPU: Towards Secure and Privacy-Preserving Knowledge Unlearning for Large Language Models
  di: Wang, Tiantong, et al.
  Pubblicazione: (2026)
• Oblivionis: A Lightweight Learning and Unlearning Framework for Federated Large Language Models
  di: Zhang, Fuyao, et al.
  Pubblicazione: (2025)
• Enhancing Federated Domain Adaptation with Multi-Domain Prototype-Based Federated Fine-Tuning
  di: Zhang, Jingyuan, et al.
  Pubblicazione: (2024)
• LoD: Loss-difference OOD Detection by Intentionally Label-Noisifying Unlabeled Wild Data
  di: Geng, Chuanxing, et al.
  Pubblicazione: (2025)
• Robust Amortized Bayesian Inference with Self-Consistency Losses on Unlabeled Data
  di: Mishra, Aayush, et al.
  Pubblicazione: (2025)