Layer-Wise Evolution of Representations in Fine-Tuned Transformers: Insights from Sparse AutoEncoders

Documenti analoghi

• Quasi-symbolic Semantic Geometry over Transformer-based Variational AutoEncoder
  di: Zhang, Yingji, et al.
  Pubblicazione: (2022)
• EBFT: Effective and Block-Wise Fine-Tuning for Sparse LLMs
  di: Guo, Song, et al.
  Pubblicazione: (2024)
• Sparse Auto-Encoders and Holism about Large Language Models
  di: Grindrod, Jumbly
  Pubblicazione: (2026)
• Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation
  di: Shu, Huizhen, et al.
  Pubblicazione: (2025)
• SafeMERGE: Preserving Safety Alignment in Fine-Tuned Large Language Models via Selective Layer-Wise Model Merging
  di: Djuhera, Aladin, et al.
  Pubblicazione: (2025)