2 papers
cs.IR2026
Sci-Surf: Navigating Scientific Literature Discovery through Human Feedback and Intelligent Summarization
Fang Guo, Qi Zhu, Rongcan Pei +3
The rapid growth of scientific publications makes it increasingly difficult for researchers to identify relevant new studies and effectively comprehend them. Existing academic disc…
cs.CL2026
Learning Less Is More: Premature Upper-Layer Attention Specialization Hurts Language Model Pretraining
Jinchang Zhu, Jindong Li, Yuwen Hao +3
A causal-decoder block is hierarchical: lower layers build the residual basis that upper layers attend over. We identify a failure mode in GPT pretraining: upper layers commit to s…