English

Sparse Activation Editing for Reliable Instruction Following in Narratives

Computation and Language 2025-05-23 v1 Artificial Intelligence Human-Computer Interaction

Abstract

Complex narrative contexts often challenge language models' ability to follow instructions, and existing benchmarks fail to capture these difficulties. To address this, we propose Concise-SAE, a training-free framework that improves instruction following by identifying and editing instruction-relevant neurons using only natural language instructions, without requiring labelled data. To thoroughly evaluate our method, we introduce FreeInstruct, a diverse and realistic benchmark of 1,212 examples that highlights the challenges of instruction following in narrative-rich settings. While initially motivated by complex narratives, Concise-SAE demonstrates state-of-the-art instruction adherence across varied tasks without compromising generation quality.

Cite

@article{arxiv.2505.16505,
  title  = {Sparse Activation Editing for Reliable Instruction Following in Narratives},
  author = {Runcong Zhao and Chengyu Cao and Qinglin Zhu and Xiucheng Lv and Shun Shao and Lin Gui and Ruifeng Xu and Yulan He},
  journal= {arXiv preprint arXiv:2505.16505},
  year   = {2025}
}
R2 v1 2026-07-01T02:31:07.422Z