Check out the new SparseUP model—ModernBERT‑based sparse embeddings that hit 58 nDCG@10 on the BEIR‑13 benchmark. Open‑source, learned sparse encoder, and ready for Linkup. Dive in! #SparseUP #BEIR13 #ModernBERT

🔗 aidailypost.com/news/new-spa...