From Unfamiliar to Familiar: Detecting Pre-training Data via Gradient Deviations in Large Language Models

Documenti analoghi

• Robust Reasoning via Dynamic Token Selection for Distribution-Aligned Self-Distillation
  di: Zhang, Ruiqi, et al.
  Pubblicazione: (2026)
• Detecting Stealthy Backdoor Samples based on Intra-class Distance for Large Language Models
  di: Chen, Jinwen, et al.
  Pubblicazione: (2025)
• AdaComp: Extractive Context Compression with Adaptive Predictor for Retrieval-Augmented Large Language Models
  di: Zhang, Qianchi, et al.
  Pubblicazione: (2024)
• MaFeRw: Query Rewriting with Multi-Aspect Feedbacks for Retrieval-Augmented Large Language Models
  di: Wang, Yujing, et al.
  Pubblicazione: (2024)
• Learning to Erase Private Knowledge from Multi-Documents for Retrieval-Augmented Large Language Models
  di: Wang, Yujing, et al.
  Pubblicazione: (2025)