# Transferring Extreme Subword Style Using Ngram Model-Based Logit Scaling

arXiv: https://arxiv.org/abs/2503.08550 · HTML: https://mlsanity.com/paper/2503.08550/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 15 references that are in the map.
- First posted on arXiv 11 Mar 2025.

## Builds on (most cited first, 8 of 8)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Hierarchical Neural Story Generation](https://mlsanity.com/paper/1805.04833/) (2018)
- [Infini-gram: Scaling Unbounded n-gram Language Models to a Trillion Tokens](https://mlsanity.com/paper/2401.17377/) (2024)
- [Style-Specific Neurons for Steering LLMs in Text Style Transfer](https://mlsanity.com/paper/2410.00593/) (2024)
- [Prompt-Based Editing for Text Style Transfer](https://mlsanity.com/paper/2301.11997/) (2023)
- [Are Large Language Models Actually Good at Text Style Transfer?](https://mlsanity.com/paper/2406.05885/) (2024)
- [Prefix-Tuning Based Unsupervised Text Style Transfer](https://mlsanity.com/paper/2310.14599/) (2023)
- [The Role of n-gram Smoothing in the Age of Neural Networks](https://mlsanity.com/paper/2403.17240/) (2024)

## Built on by (most cited first, 1 of 1)

- [Limits of n-gram Style Control for LLMs via Logit-Space Injection](https://mlsanity.com/paper/2601.16224/) (2026)

## Abstract (arXiv)

We present an ngram model-based logit scaling technique that effectively transfers extreme subword stylistic variation to large language models at inference time. We demonstrate its efficacy by tracking the perplexity of generated text with respect to the ngram interpolated and original versions of an evaluation model. Minimizing the former measure while the latter approaches the perplexity of a text produced by a target author or character lets us select a sufficient degree of adaptation while retaining fluency.