token efficiency
Token efficiency refers to the effectiveness of models, especially in natural language processing, to use fewer tokens while maintaining performance levels in tasks like understanding and generating text, thus reducing computational costs.
- ARM: Adaptive Reasoning Model
- Critical Batch Size Revisited: A Simple Empirical Approach to Large-Batch Language Model Training
- GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning
- Thoughts Are All Over the Place: On the Underthinking of Long Reasoning Models
- VaporTok: RL-Driven Adaptive Video Tokenizer with Prior & Task Awareness
- Vision Foundation Models as Effective Visual Tokenizers for Autoregressive Generation
- Vision-centric Token Compression in Large Language Model