Easy to anticipate, hard to compute: boundary dependence finds the computed outputs that entropy patching misses

arXiv:2610.11790 · cs.CL · Submitted 2026-10-08 · Read on arXiv

cs.CL

Submitted: 2026-10-08

Updated: 2026-10-08

Code: https://github.com/nicoveraz/segresearch

Terminology

• Accuracy — 13× in this paper · explained in 1 episode
• Transformer — 10× in this paper · explained in 2 episodes
• Trace — 9× in this paper · explained in 2 episodes
• Adaptation — 5× in this paper · explained in 1 episode
• Tokenization — 4× in this paper · explained in 2 episodes
• Byte Latent Transformer — 3× in this paper
• Fine-Tuning — 3× in this paper · explained in 5 episodes
• Ablation — 2× in this paper · explained in 1 episode
• Benchmark — 2× in this paper · explained in 2 episodes
• Alignment — 1× in this paper · explained in 1 episode
• Decoupling — 1× in this paper · explained in 2 episodes
• Speculative Decoding — 1× in this paper · explained in 2 episodes
• Verifier — 1× in this paper · explained in 2 episodes

Sources

• Training Verifiers to Solve Math Word Problems
• ByteSpan: Information-Driven Subword Tokenisation
• Measuring Mathematical Problem Solving With the MATH Dataset
• Dynamic Chunking for End-to-End Hierarchical Sequence Modeling
• MrT5: Dynamic Token Merging for Efficient Byte-level Language Models
• Fast Byte Latent Transformer
• Rho-1: Not All Tokens Are What You Need
• EntropyMoE: Entropy-Aware Sparse Expert Routing for Tokenizer-Free LLMs
• TokEval: A Tokenizer Evaluation Suite
• Efficient Transformers with Dynamic Token Pooling
• Byte Latent Transformer: Patches Scale Better Than Tokens
• OpenWebMath: An Open Dataset of High-Quality Mathematical Web Text
• Mixture-of-Depths: Dynamically allocating compute in transformer-based language models
• Tokenization counts: the impact of tokenization on arithmetic in frontier LLMs
• SpaceByte: Towards Deleting Tokenization from Large Language Modeling
• From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
• Scratchpad Patching: Decoupling Compute from Patch Size in Byte-Level Language Models

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering