Finding the Translation Switch: Discovering and Exploiting the Task-Initiation Features in LLMs
Xinwei Wu, Heng Liu, Xiaohu Zhao, Yuqi Ren, Linlong Xu, Longyue Wang, Deyi Xiong, Weihua Luo
Abstract
Large Language Models (LLMs) frequently exhibit strong translation abilities, even without task-specific fine-tuning. However, the internal mechanisms governing this innate capability remain largely opaque. To demystify this process, we leverage Sparse Autoencoders (SAEs) and introduce a novel framework for identifying task-specific features. Our method first recalls features that are frequently co-activated on translation inputs and then filters them for functional coherence using a PCA-based consistency metric. This framework successfully isolates a small set of "translation initiation" features. Causal interventions demonstrate that amplifying these features steers the model towards correct translation, while ablating them induces hallucinations and off-task outputs, confirming they represent a core component of the model
BibTeX
@inproceedings{aaai2026_findingthetransl,
title = {Finding the Translation Switch: Discovering and Exploiting the Task-Initiation Features in LLMs},
author = {Xinwei Wu and Heng Liu and Xiaohu Zhao and Yuqi Ren and Linlong Xu and Longyue Wang and Deyi Xiong and Weihua Luo and Kaifu Zhang},
booktitle = {AAAI 2026},
year = {2026}
}