Federated Reasoning Distillation Framework with Model Learnability-Aware Data Allocation

This research introduces LaDa, a federated reasoning distillation framework that bridges the learnability gap between large and small language models using g...

Level: advanced

By Unknown

Category: research