# Significance-Gain Pair Encoding for LLMs: A Statistical Alternative to Frequency-Based Subword Merging

arXiv: https://arxiv.org/abs/2603.19261 · HTML: https://mlsanity.com/paper/2603.19261/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 10 references that are in the map.
- First posted on arXiv 26 Feb 2026.

## Builds on (most cited first, 5 of 5)

- [Neural Machine Translation of Rare Words with Subword Units](https://mlsanity.com/paper/1508.07909/) (2015)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://mlsanity.com/paper/2105.13626/) (2021)
- [An MLP Baseline for Handwriting Recognition Using Planar Curvature and Gradient Orientation](https://mlsanity.com/paper/2508.11803/) (2025)
- [A Sobel-Gradient MLP Baseline for Handwritten Character Recognition](https://mlsanity.com/paper/2508.11902/) (2025)

## Built on by (most cited first, 2 of 2)

- [StepCache: Step-Level Reuse with Lightweight Verification and Selective Patching for LLM Serving](https://mlsanity.com/paper/2603.28795/) (2026)
- [MDL-Calibrated Significance-Gain Pair Encoding: Replication-Aware Automatic Stopping for Subword Tokenization](https://mlsanity.com/paper/2609.31705/) (2026)

## Abstract (arXiv)

Subword tokenization is a key design choice for modern language models, including large language models (LLMs), with byte- and character-level BPE serving as a widely used baseline. Standard BPE selects merges by raw pair frequency, which favors compression but can conflate true adjacency cohesion with pairs that are frequent due to high marginal counts. This paper introduces Significance-Gain BPE, a drop-in alternative merge criterion that measures cohesion via a z-statistic under an independence null model and combines it with an explicit compression-aware gain term. Significance-Gain BPE is evaluated on WikiText-103 (raw) character slices using a small causal Transformer language model, reporting both token-dependent perplexity and the tokenizer-invariant metric bits per character (BPC). At a representative operating point, Significance-Gain BPE reduces validation and test perplexity by 13% and 12%, respectively, and improves validation and test BPC by about 0.9 to 1.0%. A vocabulary-size sweep further shows lower BPC in most closest-compression comparisons, suggesting that statistically grounded merge selection can improve predictive efficiency per unit of raw text across a range of compression regimes.