3 papers
cs.CL2025
ChuLo: Chunk-Level Key Information Representation for Long Document Understanding
Yan Li, Soyeon Caren Han, Yue Dai +1
Transformer-based models have achieved remarkable success in various Natural Language Processing (NLP) tasks, yet their ability to handle long documents is constrained by computati…
cs.CL2025
A Training-Free Length Extrapolation Approach for LLMs: Greedy Attention Logit Interpolation (GALI)
Yan Li, Tianyi Zhang, Zechuan Li +1
Transformer-based Large Language Models (LLMs) struggle with inputs exceeding their training context window due to positional out-of-distribution (O.O.D.) issues that disrupt atten…
cs.CL2025
MIDAS: Multi-level Intent, Domain, And Slot Knowledge Distillation for Multi-turn NLU
Yan Li, So-Eon Kim, Seong-Bae Park +1
Although Large Language Models (LLMs) can generate coherent text, they often struggle to recognise user intent behind queries. In contrast, Natural Language Understanding (NLU) mod…