CURE: Controlled Unlearning for Robust Embeddings -- Mitigating Conceptual Shortcuts in Pre-Trained Language Models

Documenti analoghi

• RAZOR: Sharpening Knowledge by Cutting Bias with Unsupervised Text Rewriting
  di: Yang, Shuo, et al.
  Pubblicazione: (2024)
• Doubling Your Data in Minutes: Ultra-fast Tabular Data Generation via LLM-Induced Dependency Graphs
  di: Yang, Shuo, et al.
  Pubblicazione: (2025)
• Active Tabular Augmentation via Policy-Guided Diffusion Inpainting
  di: Zhang, Zheyu, et al.
  Pubblicazione: (2026)
• SCISSOR: Mitigating Semantic Bias through Cluster-Aware Siamese Networks for Robust Classification
  di: Yang, Shuo, et al.
  Pubblicazione: (2025)
• From Confidence to Collapse in LLM Factual Robustness
  di: Fastowski, Alina, et al.
  Pubblicazione: (2025)