Peek2: Regex-free Byte-level Byte-Pair Encoding Pretokenizer for LLM Inference on Edge Devices

arXiv:2601.05833v2 Announce Type: replace Abstract: Pretokenization is a crucial, sequential pass in Byte-level BPE tokenizers, yet little work has been done to optimize it for edge-side inference. Our proposed new implementation, Peek2, serves as a drop-in replacement for cl100k-like pretokenizers used in GPT-3, LLaMa-3, and Qwen-2.5. After breaking down and analyzing the logic of the original cl100k pretokenizer, we introduced a new pretokenization algorithm with linear time complexity and constant, trivial memory usage, suited for edge scenarios. Test results show that it increases microbenchmarking throughput by up to 2.48imes and delivers a 1.14imes improvement in overall throughput across the entire Byte-level BPE encoding process, depending on the dataset, while providing identical results as the baseline Regex-based tokenizer.

Source: arXiv cs.CL | 2026-05-04