PromptDistill: Query-based Selective Token Retention in Intermediate Layers for Efficient Large Language Model Inference

Documenti analoghi

• LLMs Can't Handle Peer Pressure: Crumbling under Multi-Agent Social Interactions
  di: Song, Maojia, et al.
  Pubblicazione: (2025)
• Error Typing for Smarter Rewards: Improving Process Reward Models with Error-Aware Hierarchical Supervision
  di: Pala, Tej Deep, et al.
  Pubblicazione: (2025)
• DELLA-Merging: Reducing Interference in Model Merging through Magnitude-Based Sampling
  di: Deep, Pala Tej, et al.
  Pubblicazione: (2024)
• Ferret: Faster and Effective Automated Red Teaming with Reward-Based Scoring Technique
  di: Pala, Tej Deep, et al.
  Pubblicazione: (2024)
• Lessons from Training Grounded LLMs with Verifiable Rewards
  di: Sim, Shang Hong, et al.
  Pubblicazione: (2025)