ICML 2025PastNLP
Tokenization Workshop
TokShop
- Submission deadline
- Jun 1, 2025, 11:59 UTCimported from OpenReview — check the website for extensions
- Submission portal
- OpenReview
- Notes
- Topics were auto-suggested and may be imprecise — edits welcome.
Accepted papers (37)
Fetched from OpenReview (v2) on 2026-06-10.
Adversarial Tokenization
BPE Stays on SCRIPT: Structured Encoding for Robust Multilingual Pretokenization
Broken Tokens? Your Language Model can Secretly Handle Non-Canonical Tokenizations
Byte Latent Transformer: Patches Scale Better Than Tokens
Byte-level Tokenizers Unavoidably Enable LLMs to Generate Ill-formed UTF-8
ByteSpan: Information-Driven Subword Tokenisation
Canonical Autoregressive Generation
CAT: Content-Adaptive Image Tokenization
Causal Estimation of Tokenisation Bias
Conditional Unigram Tokenization with Parallel Data
Contextual morphologically-guided tokenization for pretrained Latin BERT models
Continuous Autoregressive Generation with Mixture of Gaussians
Continuous Chain of Thought Enables Parallel Exploration and Reasoning
Discrete JEPA: Learning Discrete Token Representations without Reconstruction
Entropy-Driven Pre-Tokenization for Byte-Pair Encoding
Evaluating Morphological Alignment of Tokenizers in 70 Languages
FLEXITOKENS: Flexible Tokenization for Evolving Language Models
GeneticBPE: Motif-Preserving Tokenization for Robust miRNA Modeling
HH-Codec: High Compression High-fidelity Discrete Neural Codec for Spoken Language Modeling
How Much is Enough? The Diminishing Returns of Tokenization Training Data
How Tokenization Limits Phonological Knowledge Representation in Language Models and How to Improve Them
InCa and InDia: Inline Casing and Diacritization Preprocessing For Robust-to-Noise Tokenization and Interpretability
Is Your LLM Overcharging You? Tokenization, Transparency, and Incentives
MorphTok: Morphologically Grounded Tokenization for Indic languages
Motion-Focused Tokenization for Source-Free Video Domain Adaptation
One-D-Piece: Image Tokenizer Meets Quality-Controllable Compression
Overcoming Vocabulary Constraints with Pixel-level Fallback
Pitfalls, Subtleties, and Techniques in Automata-Based Subword-Level Constrained Generation
QuickMerge++: Token Merging with Autoregressive Prior
Robust Noise Attenuation via Adaptive Pooling of Transformer Outputs
Sampling from Your Language Model One Byte at a Time
SuperBPE: Space Travel for Language Models
Tokenisation is NP-Complete
Tokenizing Nonverbal Communication in Salsa Dance
Watermarking Autoregressive Image Generation
You Only Train Once: Efficient Tokenizer Selection for Arithmetic in Language Models
zip2zip: Inference-Time Adaptive Vocabularies for Language Models via Token Compression