arXiv:2606. 15144v1 Announce Type: cross Abstract: Large language models (LLMs) process text as sequences of subword tokens, which can obscure the character-level and morphological structure that underlies word formation.
arXiv:2607. 23319v1 Announce Type: cross Abstract: Standard subword tokenization algorithms such as Byte-Pair Encoding (BPE) and SentencePiece are trained predominantly on modern language corpora and produce inefficient segmentations when applied to classical Indian languages.
arXiv:2606. 15521v1 Announce Type: cross Abstract: Tokenization introduces representational redundancy: under a fixed token vocabulary, every byte string admits many valid token encodings, or segmentations, that decode to the same surface string.
arXiv:2606. 25987v1 Announce Type: cross Abstract: Large language models (LLMs) attain remarkable surface fluency on code, yet they neither formally guarantee the syntactic validity of their output nor leverage the hierarchical structure defining the target language.
CWoMP (Contrastive Word‑Morpheme Pretraining) is a new approach for generating interlinear glossed text that treats morphemes as atomic form‑meaning units with learned representations. It uses a contrastively trained encoder to align words in context with their constituent morphemes in a shared embedding space, and an autoregressive decoder that retrieves morpheme sequences from a mutable lexicon of these embeddings. The method yields interpretable predictions grounded in lexicon entries and allows users to improve results at inference time by expanding the lexicon without retraining, achieving superior performance and efficiency on diverse low‑resource languages, especially in extremely low‑resource settings.
TokEval is a tokenizer evaluation suite that introduces metrics beyond traditional fertility and compression rate, capturing linguistically and structurally meaningful properties such as UTF-8 character boundary integrity and digit place-value alignment for mathematics. The authors validate these metrics by conducting controlled language model pretraining experiments that vary tokenizer training data, pretokenization strategy, and training algorithm, then evaluate the models on bits-per-byte and benchmarks covering linguistic understanding, mathematical reasoning, and code generation. Results show that information-theoretic metrics predict language modeling performance, while structure-sensitive metrics correlate with task accuracy, suggesting TokEval can guide tokenizer selection more principledly.
The paper introduces two new tokenisation algorithms—BottomUpLL and TopDownComp—to systematically explore the 2x2 design space defined by optimisation objective (compression vs. log‑likelihood) and search procedure (bottom‑up merging vs. top‑down pruning). Experiments across model sizes, vocabularies, and domains show that the search procedure, rather than the objective, consistently yields lower bits‑per‑byte, while no clear pattern emerges on the BLiMP benchmark. These findings clarify how tokeniser design choices influence language‑model performance and provide guidance for constructing tokenisers more principledly.
arXiv:2511. 20849v2 Announce Type: replace-cross Abstract: We introduce a new tokenizer for language models that minimizes the average tokens per character, thereby reducing the number of tokens needed to represent text during training and to generate text during inference.
TokEval is a tokenizer evaluation suite that introduces metrics beyond traditional fertility and compression rate, capturing linguistically and structurally meaningful properties such as UTF‑8 character boundary integrity and digit place‑value alignment for mathematics. The authors validate these metrics by pretraining language models with varied tokenizers and measuring downstream performance on bits‑per‑byte and benchmarks covering linguistic understanding, mathematical reasoning, and code generation. Their results show that information‑theoretic metrics predict language modeling performance, while structure‑sensitive metrics correlate with task accuracy, suggesting TokEval can guide tokenizer selection more principledly.
arXiv:2506. 15138v2 Announce Type: replace-cross Abstract: Tokenization directly affects the inference efficiency of large language models, since fragmented tokenization increases sequence length and generation cost.
The paper examines how new vocabulary tokens are added to language models for generative recommendation tasks. It shows that the common practice of initializing these tokens as the mean of existing embeddings collapses them into a degenerate subspace, hindering fine‑tuning. The authors propose Grounded Token Initialization (GTI), which places new tokens at semantically meaningful positions in the pretrained embedding space using linguistic supervision, and demonstrate that GTI outperforms mean initialization and other adaptation methods across several benchmarks.
arXiv:2608. 03494v1 Announce Type: cross Abstract: Vocabulary extension is an efficient way to adapt pretrained large language models (LLMs) to new languages, but the initialization of newly added token embeddings can strongly affect continued pre-training (CPT) efficiency.