一作ACL 2026Poster· Natural Language Processing
SPEAK : Spiking Neurons as an Entropy-Aware Tokenizer for Large Language Models
Ming Chen · Wenyao Li · Chao Liang · Shi Gu · Peng Lin · De Ma · Huajin Tang · Qian Zheng · Gang Pan
摘要
Tokenizers play a critical role in large language model studies. Despite recent advances, existing tokenizers fail to explicitly leverage historical tokenization results when making subsequent token decisions, nor do they selectively utilize such history based on contextual relevance. We propose...