Heads up! Large Language Models Can Perform Tasks Without Your Instruction via Selective Attention Head Masking

Senyu Han, Hongchuan Zeng, Kai Yu, Lu Chen
2/14/2026

No official code found. Best implementation: OpenDFM/HeadsUp (3 stars, last commit 2025-06-20).

Semantic Scholar

Code Implementations(1)

OpenDFM/HeadsUpCommunity69%

[ICML 2025] Codes for the paper "Heads up! Large Language Models Can Perform Tasks Without Your Instruction via Selective Attention Head Masking"

31Shell, PythonJan 14, 20251 years ago
interpretable-deep-learningllm

Similar Papers

Privacy-Preserving Embedding via Look-up Table Evaluation with Fully Homomorphic Encryption

92% match2026

PoseIRM: Enhance 3D Human Pose Estimation on Unseen Camera Settings via Invariant Risk Minimization

84% match1 repo32026

Clustered Federated Learning via Gradient-based Partitioning

76% match1 repo152026

Open-Set Domain Adaptation via Joint Error Based Multi-class Positive and Unlabeled Learning

71% match2026

Multi-level Protein Structure Pre-training via Prompt Learning

70% match2026

ChemAgent: Self-updating Memories in Large Language Models Improves Chemical Reasoning

63% match2026

Cite this paper

@article{han2026heads,
  title  = {Heads up! Large Language Models Can Perform Tasks Without Your Instruction via Selective Attention Head Masking},
  author = {Senyu Han and Hongchuan Zeng and Kai Yu and Lu Chen},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:30eac4552ccae8985f303887c8464e76002fa5ff},
  journal = {ICML 2025 2025}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/30eac4552ccae8985f303887c8464e76002fa5ff.svg)](https://codeofpaper.com/paper/30eac4552ccae8985f303887c8464e76002fa5ff)

Discussion