Yilun Liu

Paper here: aclanthology.org/2026.findings-…

We're delighted to present our work "Parameter-Efficient Routed Fine-Tuning: Mixture-of-Experts Demands Mixture of Adaptation Modules" @ EACL'26! 🎈 We analyze core dynamics when PEFTing MoE LLMs, introduce a unified framework & a family of strategies. Thanks to all coauthors!

🏮Want to classify text more accurately, efficiently, and interpretably? Today at #ACL2024, we present our method SPIN which uses not just the last layer of a neural network, but learns how to leverage the entire internal structure for best performance. 👇

Thanks to all co-authors for the amazing work 😎 See you in Bangkok!  #ACL #ACL2024NLP #LLMs #AI #NLProc #UofT #TU_Muenchen

2⃣ We conduct extensive experiments to demonstrate SPIN’s superior performance, improved training and inference efficiency, and enhanced intrinsic and post-hoc interpretability in text classification.

🌟 We are delighted to announce that our paper "🐚SPIN: Sparsifying and Integrating Internal Neurons in Large Language Models for Text Classification" (w/ @DifanJ2000, @lilvjosephtang, @DanielMatter96, @JurgenPfeffer, @ashton1anderson) has been accepted at #ACL2024 findings!🧵👇