Finding the Translation Switch: Discovering and Exploiting the Task-Initiation Features in LLMs

TL;DR

This paper uncovers the internal translation initiation features in large language models using autoencoders and causal interventions, and applies this insight to improve fine-tuning efficiency and robustness.

Contribution

It introduces a novel framework for identifying translation initiation features in LLMs and demonstrates their transferability and utility in enhancing fine-tuning strategies.

Findings

Isolated core translation initiation features in LLMs.

Amplifying these features improves translation accuracy.

Focusing on mechanistically hard samples enhances data efficiency.

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's innate translation competency. Moving from analysis to application, we leverage this mechanistic insight to propose a new data selection strategy for efficient fine-tuning. Specifically, we prioritize training on **mechanistically hard** samples-those that fail to naturally activate the translation initiation features. Experiments show this approach significantly improves data efficiency and suppresses hallucinations. Furthermore, we find these mechanisms are transferable to larger models of the same family. Our work not only decodes a core component of the translation mechanism in LLMs but also provides a blueprint for using internal model mechanism to create more robust and efficient models. The codes are available at https://github.com/flamewei123/AAAI26-translation-Initiation-Features.