# Steering into New Embedding Spaces: Analyzing Cross-Lingual Alignment Induced by Model Interventions in Multilingual Language Models

arXiv: https://arxiv.org/abs/2502.15639 · HTML: https://mlsanity.com/paper/2502.15639/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 44 references that are in the map.
- First posted on arXiv 21 Feb 2025.

## Builds on (most cited first, 27 of 27)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [BLOOM: A 176B-Parameter Open-Access Multilingual Language Model](https://mlsanity.com/paper/2211.05100/) (2022)
- [No Language Left Behind: Scaling Human-Centered Machine Translation](https://mlsanity.com/paper/2207.04672/) (2022)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Cross-lingual Language Model Pretraining](https://mlsanity.com/paper/1901.07291/) (2019)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization](https://mlsanity.com/paper/2003.11080/) (2020)
- [Text Summarization with Pretrained Encoders](https://mlsanity.com/paper/1908.08345/) (2019)
- [COMET: Commonsense Transformers for Automatic Knowledge Graph Construction](https://mlsanity.com/paper/1906.05317/) (2019)
- [Do Llamas Work in English? On the Latent Language of Multilingual Transformers](https://mlsanity.com/paper/2402.10588/) (2024)
- [Unsupervised Machine Translation Using Monolingual Corpora Only](https://mlsanity.com/paper/1711.00043/) (2017)
- [Not All Languages Are Created Equal in LLMs: Improving Multilingual Capability by Cross-Lingual-Thought Prompting](https://mlsanity.com/paper/2305.07004/) (2023)
- [LLaMA Beyond English: An Empirical Study on Language Capability Transfer](https://mlsanity.com/paper/2401.01055/) (2024)
- [On the Multilingual Ability of Decoder-based Pre-trained Language Models: Finding and Controlling Language-Specific Neurons](https://mlsanity.com/paper/2404.02431/) (2024)
- [Multilingual LLMs are Better Cross-lingual In-context Learners with Alignment](https://mlsanity.com/paper/2305.05940/) (2023)
- [Controlling Language and Diffusion Models by Transporting Activations](https://mlsanity.com/paper/2410.23054/) (2024)
- [Improving In-context Learning of Multilingual Generative Language Models with Cross-lingual Alignment](https://mlsanity.com/paper/2311.08089/) (2023)
- [Whispering Experts: Neural Interventions for Toxicity Mitigation in Language Models](https://mlsanity.com/paper/2407.12824/) (2024)
- [Measuring and Reducing LLM Hallucination without Gold-Standard Answers](https://mlsanity.com/paper/2402.10412/) (2024)
- [Geographic and Geopolitical Biases of Language Models](https://mlsanity.com/paper/2212.10408/) (2022)
- [DICT-MLM: Improved Multilingual Pre-Training using Bilingual Dictionaries](https://mlsanity.com/paper/2010.12566/) (2020)
- [How Transliterations Improve Crosslingual Alignment](https://mlsanity.com/paper/2409.17326/) (2024)
- [The Impact of Cross-Lingual Adjustment of Contextual Word Representations on Zero-Shot Transfer](https://mlsanity.com/paper/2204.06457/) (2022)

## Built on by (most cited first, 11 of 11)

- [Can you map it to English? The Role of Cross-Lingual Alignment in Multilingual Performance of LLMs](https://mlsanity.com/paper/2504.09378/) (2025)
- [Cross-Lingual Activation Steering for Multilingual Language Models](https://mlsanity.com/paper/2601.16390/) (2026)
- [Bias Beyond Borders: Political Ideology Evaluation and Steering in Multilingual LLMs](https://mlsanity.com/paper/2601.23001/) (2026)
- [Language Lives in Sparse Dimensions: Toward Interpretable and Efficient Multilingual Control for Large Language Models](https://mlsanity.com/paper/2510.07213/) (2025)
- [Sparse Subnetwork Enhancement for Underrepresented Languages in Large Language Models](https://mlsanity.com/paper/2510.13580/) (2025)
- [Discriminating Form and Meaning in Multilingual Models with Minimal-Pair ABX Tasks](https://mlsanity.com/paper/2505.17747/) (2025)
- [Multilinguality as Sense Adaptation](https://mlsanity.com/paper/2601.10310/) (2026)
- [A Survey of Toxicity Detection and Mitigation Strategies for Multilingual Language Models](https://mlsanity.com/paper/2606.25380/) (2026)
- [Why Vision Fails as a Universal Bridge: Rectifying Modality Asynchrony in Multilingual MLLMs](https://mlsanity.com/paper/2608.15085/) (2026)
- [Target-Language Generation in Multilingual Models: Activation Steering and Optimal Control](https://mlsanity.com/paper/2609.16967/) (2026)
- [Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders](https://mlsanity.com/paper/2610.18526/) (2026)

## Abstract (arXiv)

Aligned representations across languages is a desired property in multilingual large language models (mLLMs), as alignment can improve performance in cross-lingual tasks. Typically alignment requires fine-tuning a model, which is computationally expensive, and sizable language data, which often may not be available. A data-efficient alternative to fine-tuning is model interventions -- a method for manipulating model activations to steer generation into the desired direction. We analyze the effect of a popular intervention (finding experts) on the alignment of cross-lingual representations in mLLMs. We identify the neurons to manipulate for a given language and introspect the embedding space of mLLMs pre- and post-manipulation. We show that modifying the mLLM's activations changes its embedding space such that cross-lingual alignment is enhanced. Further, we show that the changes to the embedding space translate into improved downstream performance on retrieval tasks, with up to 2x improvements in top-1 accuracy on cross-lingual retrieval.