On the Multilingual Ability of Decoder-based Pre-trained Language Models: Finding and Controlling Language-Specific Neurons

Documenti analoghi

• Inconsistent Tokenizations Cause Language Models to be Perplexed by Japanese Grammar
  di: Gambardella, Andrew, et al.
  Pubblicazione: (2025)
• Which Programming Language and What Features at Pre-training Stage Affect Downstream Logical Inference Performance?
  di: Uchiyama, Fumiya, et al.
  Pubblicazione: (2024)
• Automated Refinement of Essay Scoring Rubrics for Language Models via Reflect-and-Revise
  di: Harada, Keno, et al.
  Pubblicazione: (2025)
• Semantic Token Clustering for Efficient Uncertainty Quantification in Large Language Models
  di: Cao, Qi, et al.
  Pubblicazione: (2026)
• NeuronMoE: Neuron-Guided Mixture-of-Experts for Efficient Multilingual LLM Extension
  di: Li, Rongzhi, et al.
  Pubblicazione: (2026)