🧑🔬 Kushal Tirumala — papers of 2023 — Elamur Research2 papers · all years of Kushal TirumalaSemDeDup: Data-efficient learning at web-scale through semantic deduplicationarXiv:2303.09540Amro Abbas, Kushal Tirumala, Dániel Simig, et al.2023 · 108 citationsD4: Improving LLM Pretraining via Document De-Duplication and DiversificationConference on Neural Information Processing Systems (NeurIPS)Kushal Tirumala, Daniel Simig, Armen Aghajanyan, et al.2023 · 81 citationsThis page in Markdown → https://md.research.elamur.ai/free/en/authors/kushal-tirumala/year/2023/