# Steering Large Language Models for Machine Translation Personalization

arXiv: https://arxiv.org/abs/2505.16612 · HTML: https://mlsanity.com/paper/2505.16612/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Activation steering methods, one of 13 directions in Mechanistic interpretability.
- Builds on 35 of its 68 references that are in the map.
- First posted on arXiv 22 May 2025.

## Placed in

- Mechanistic interpretability / [Activation steering methods](https://mlsanity.com/mechinterp/activation-steering-methods/)

## Builds on (most cited first, 30 of 35)

- [Unsupervised Cross-lingual Representation Learning at Scale](https://mlsanity.com/paper/1911.02116/) (2019)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [COMET: A Neural Framework for MT Evaluation](https://mlsanity.com/paper/2009.09025/) (2020)
- [Refusal in Language Models Is Mediated by a Single Direction](https://mlsanity.com/paper/2406.11717/) (2024)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [Probing Classifiers: Promises, Shortcomings, and Advances](https://mlsanity.com/paper/2102.12452/) (2021)
- [Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models](https://mlsanity.com/paper/2403.19647/) (2024)
- [Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://mlsanity.com/paper/2408.05147/) (2024)
- [In-Context Learning Creates Task Vectors](https://mlsanity.com/paper/2310.15916/) (2023)
- [AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders](https://mlsanity.com/paper/2501.17148/) (2025)
- [Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models](https://mlsanity.com/paper/2411.14257/) (2024)
- [Transformer visualization via dictionary learning: contextualized embedding as a linear superposition of transformer factors](https://mlsanity.com/paper/2103.15949/) (2021)
- [Getting Gender Right in Neural Machine Translation](https://mlsanity.com/paper/1909.05088/) (2019)
- [Improving Steering Vectors by Targeting Sparse Autoencoder Features](https://mlsanity.com/paper/2411.02193/) (2024)
- [Reducing Gender Bias in Neural Machine Translation as a Domain Adaptation Problem](https://mlsanity.com/paper/2004.04498/) (2020)
- [Large language models effectively leverage document-level context for literary translation, but critical errors persist](https://mlsanity.com/paper/2304.03245/) (2023)
- [Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering](https://mlsanity.com/paper/2410.15999/) (2024)
- [(Perhaps) Beyond Human Translation: Harnessing Multi-Agent Collaboration for Translating Ultra-Long Literary Texts](https://mlsanity.com/paper/2405.11804/) (2024)
- [Personalized Machine Translation: Preserving Original Author Traits](https://mlsanity.com/paper/1610.05461/) (2016)
- [Exploring Document-Level Literary Machine Translation with Parallel Paragraphs from World Literature](https://mlsanity.com/paper/2210.14250/) (2022)
- [Multi-property Steering of Large Language Models with Dynamic Activation Composition](https://mlsanity.com/paper/2406.17563/) (2024)
- [Translating Step-by-Step: Decomposing the Translation Process for Improved Translation Quality of Long-Form Texts](https://mlsanity.com/paper/2409.06790/) (2024)
- [Extreme Adaptation for Personalized Neural Machine Translation](https://mlsanity.com/paper/1805.01817/) (2018)
- [Using natural language prompts for machine translation](https://mlsanity.com/paper/2202.11822/) (2022)
- [Findings of the WMT 2023 Shared Task on Discourse-Level Literary Translation: A Fresh Orb in the Cosmos of LLMs](https://mlsanity.com/paper/2311.03127/) (2023)
- [CoCoA-MT: A Dataset and Benchmark for Contrastive Controlled MT with Application to Formality](https://mlsanity.com/paper/2205.04022/) (2022)
- [Findings of the WMT 2024 Shared Task on Discourse-Level Literary Translation](https://mlsanity.com/paper/2412.11732/) (2024)

## Built on by (most cited first, 3 of 3)

- [Beyond a Global Norm: Personalizing Toxicity Sensitivity in Language Models Without Retraining](https://mlsanity.com/paper/2607.23175/) (2026)
- [A Cross-lingual Comparison of Human and Classification Model Entrainment Behavior in Code-switched Speech Settings](https://mlsanity.com/paper/2607.25202/) (2026)
- [Model vs. User: Whose Values Prevail? Inference-Time Value Alignment in Multi-Turn Conversations](https://mlsanity.com/paper/2610.19817/) (2026)

## Abstract (arXiv)

Large language models have simplified the production of personalized translations reflecting predefined stylistic constraints. However, these systems still struggle when stylistic requirements are implicitly represented by a set of examples, such as texts produced by a specific human translator. In this work, we explore various strategies for personalizing automatically generated translations when few examples are available, with a focus on the challenging domain of literary translation. We begin by determining the feasibility of the task and how style information is encoded within model representations. Then, we evaluate various prompting strategies and inference-time interventions for steering model generations towards a personalized style, with a particular focus on contrastive steering with sparse autoencoder (SAE) latents to identify salient personalization properties. We demonstrate that contrastive SAE steering yields robust style conditioning and translation quality, resulting in higher inference-time computational efficiency than prompting approaches. We further examine the impact of steering on model activations, finding that layers encoding personalization properties are impacted similarly by prompting and SAE steering, suggesting a similar mechanism at play.