VOCABTRIM: Vocabulary Pruning for Efficient Speculative Decoding in LLMs

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

1. 1Direct Alignment of Draft Model for Speculative Decoding with Chat-Fine-Tuned LLMs2024
2. 2S2D: Sorted Speculative Decoding For More Efficient Deployment of Nested Large Language Models2024
3. 3Recursive Speculative Decoding: Accelerating LLM Inference via Sampling Without Replacement2024
4. 4SpecVLM: Fast Speculative Decoding in Vision-Language Models2025
5. 5SpecVLM: Enhancing Speculative Decoding of Video LLMs via Verifier-Guided Token Pruning2025