FlashHead: Efficient Drop-In Replacement for the Classification Head in Language Model Inference

Documenti analoghi

• HeadRank: Decoding-Free Passage Reranking via Preference-Aligned Attention Heads
  di: Wang, Juyuan, et al.
  Pubblicazione: (2026)
• Learning to Route Queries to Heads for Attention-based Re-ranking with Large Language Models
  di: Tian, Yuxing, et al.
  Pubblicazione: (2026)
• Cutting Off the Head Ends the Conflict: A Mechanism for Interpreting and Mitigating Knowledge Conflicts in Language Models
  di: Jin, Zhuoran, et al.
  Pubblicazione: (2024)
• MAGE: Multi-Head Attention Guided Embeddings for Low Resource Sentiment Classification
  di: Vashisht, Varun, et al.
  Pubblicazione: (2025)
• Multi-Head RAG: Solving Multi-Aspect Problems with LLMs
  di: Besta, Maciej, et al.
  Pubblicazione: (2024)