4 papers
From Self-Attention to Connection Laplacian: A Unified Operator View of Transformers
Binbin Lin, Wei Chen, Yalun Li +3
Self-attention is a ubiquitous primitive in modern sequence models, yet its operator-level geometry is only partially understood. We view a token sequence as a vector field over th…
Leveraging Submodule Linearity Enhances Task Arithmetic Performance in LLMs
Rui Dai, Sile Hu, Xu Shen +3
Task arithmetic is a straightforward yet highly effective strategy for model merging, enabling the resultant model to exhibit multi-task capabilities. Recent research indicates tha…
From Yes-Men to Truth-Tellers: Addressing Sycophancy in Large Language Models with Pinpoint Tuning
Wei Chen, Zhen Huang, Liang Xie +9
Large Language Models (LLMs) tend to prioritize adherence to user prompts over providing veracious responses, leading to the sycophancy issue. When challenged by users, LLMs tend t…
Enhancing Multiple Dimensions of Trustworthiness in LLMs via Sparse Activation Control
Yuxin Xiao, Chaoqun Wan, Yonggang Zhang +5
As the development and application of Large Language Models (LLMs) continue to advance rapidly, enhancing their trustworthiness and aligning them with human preferences has become…