Keshuang Li — accepted AI-conference papers

2026

Spectral Attention Steering for Prompt Highlighting

Weixian Waylon Li, Yuchen Niu, Yongxin Yang, Keshuang Li, Tiejun Ma, Shay B Cohen

ICLR 2026poster

Steering a large language model's attention towards user-specified highlighted text is a critical capability. Existing prompt highlighting methods are incompatible with modern efficient attention mechanisms like Flash Attention due to their reliance on post-hoc matrix editing. We introduce Spectral…

Cited by 0 Source ↗code ↗Scholar ↗

2025

From an LLM Swarm to a PDDL-empowered Hive: Planning Self-executed Instructions in a Multi-modal Jungle

Kaustubh Vyas, Damien Graux, Yijun Yang, Sebastien Montella, Chenxin Diao, Wendi Zhou +5

ICLR 2025poster

In response to the call for agent-based solutions that leverage the ever-increasing capabilities of the deep models' ecosystem, we introduce a comprehensive solution for selecting appropriate models and subsequently planning a set of atomic actions to satisfy the end-users' instructions. Our system…

Cited by 0 Source ↗PDF ↗Scholar ↗