Pith. sign in

arXiv preprint arXiv:2410.23111 , year=

Exploring Gradient Subspaces: Addressing, Overcoming LoRA's Limitations in Federated Fine-Tuning of Large Language Models , author= · arXiv 2410.23111

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
read on arXiv browse 1 citing papers

fields

cs.LG 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

Subspace-Constrained Federated Learning with Low-Rank Adaptation

cs.LG · 2026-06-21 · unverdicted · novelty 5.0

Subspace-Reg federated LoRA yields higher accuracy and near-perfect subspace overlap (~0.9999) on RoBERTa-large in 10-client non-IID HellaSwag experiments versus FedAvg and SVD baselines, but FedAvg wins on SmolLM-360M.

citing papers explorer

Showing 1 of 1 citing paper.

  • Subspace-Constrained Federated Learning with Low-Rank Adaptation cs.LG · 2026-06-21 · unverdicted · none · ref 12

    Subspace-Reg federated LoRA yields higher accuracy and near-perfect subspace overlap (~0.9999) on RoBERTa-large in 10-client non-IID HellaSwag experiments versus FedAvg and SVD baselines, but FedAvg wins on SmolLM-360M.