Compositional Multilingual and Behavioral Attribute Steering

Abstract

This study examines the compositionality of steering vectors for language and behavioral control in large language models. Focusing on language, jailbreak, and conciseness, we investigate whether additive, training-free composition of attribute steering vectors can preserve the intended steering effect of each attribute, across four instruction-tuned models from two model families and two size scales. We find that single-attribute steering is reliable for all three attributes, but only within an

Categories

Framework mappings

Suggested from the entry's categories.

@misc{kang2026compositional,
  title = {{Compositional Multilingual and Behavioral Attribute Steering}},
  author = {Hyun Gu Kang and Daniil Gurgurov and Tanja Baeumel and Josef van Genabith and Simon Ostermann},
  year = {2026},
  month = sep,
  eprint = {2609.08410},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.08410}
}