3

Humanity's Last Exam
Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are …
More than 600 authors including, Donato Crisostomi, Emanuele Rodolà
Cite URL GitHub arXiv
Humanity's Last Exam
ATM: Improving Model Merging by Alternating Tuning and Merging
Model merging has recently emerged as a cost-efficient paradigm for multi-task learning. Among current approaches, task arithmetic …
Luca Zhou, Daniele Solombrino, Donato Crisostomi, Maria Sofia Bucarelli, Fabrizio Silvestri, Emanuele Rodolà
Cite arXiv GitHub
ATM: Improving Model Merging by Alternating Tuning and Merging