From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization

Documenti analoghi

• Task-Stratified Knowledge Scaling Laws for Post-Training Quantized Large Language Models
  di: Zhou, Chenxi, et al.
  Pubblicazione: (2025)
• Knowledge in Superposition: Unveiling the Failures of Lifelong Knowledge Editing for Large Language Models
  di: Hu, Chenhui, et al.
  Pubblicazione: (2024)
• Judging Against the Reference: Uncovering Knowledge-Driven Failures in LLM-Judges on QA Evaluation
  di: Lee, Dongryeol, et al.
  Pubblicazione: (2026)
• Catastrophic Failure of LLM Unlearning via Quantization
  di: Zhang, Zhiwei, et al.
  Pubblicazione: (2024)
• The Knowledge Microscope: Features as Better Analytical Lenses than Neurons
  di: Chen, Yuheng, et al.
  Pubblicazione: (2025)