Pilot corpus only
This score is computed over theSindex pilot corpus and does not cover the full scientific literature. Scores are relative to papers we have ingested — papers, authors, and institutions outside the pilot are not represented. Methodology.
Driven by 1 paper. Top paper: “Taming Latency-Memory Trade-Off in MoE-Based LLM Serving via Fine-Grained Expert Offloading”.
Cui, Xingqi, Hong Zhang, Hao Wang, Hanfei Yu