AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

MobiEdit: Resource-efficient Knowledge Editing for Personalized On-device LLMs

Zhenyan Lu, Daliang Xu, Dongqi Cai, Zexi Li, Wei Liu, Jian Luan, Fangming Liu, Shangguang Wang, Mengwei Xu

ICLR 2026 · Large language models › Retrieval and knowledge · approach: Knowledge-augmented question answering · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add early stopping controller to zeroth-order editing (ablation)MobiEdit zo · ZsREediting time: over 40% less timeImproved
add early stopping controller to zeroth-order editing (ablation)MobiEdit zo · ZsREedit successNo clear effect
add prefix cache on top of early stopping (ablation)MobiEdit · ZsREediting time: another 20-30% fasterImproved
combine early stopping and prefix cache (ablation)basic zeroth-order editing · ZsREediting time: over 4000 seconds -> nearly one-thirdImproved
use quantized forward-only editing on NPU instead of ROME on CPUROME · ZsRE (K60 Pro)editing time (s): 4543.78 -> 1902.88Improved
use MobiEdit instead of ROME/MEMIT/AlphaEditROME · ZsREmemory (GB): 46.14 -> 6.20Improved
use MobiEdit instead of WISEWISE · ZsRE (K60 Pro)energy per edit (J): 0.63 -> 0.023Improved
use zeroth-order estimates instead of backpropagationBP-based editing · ZsRE and CounterFactnumber of steps: 20x more stepsHurt
use MobiEdit instead of MEMIT and AlphaEditMEMIT/AlphaEdit · ZsRE and CounterFactedit success: 13.9 reduction in edit successHurt
reuse stale prefix activations via prefix cache (ablation)MobiEdit · ZsREcosine similarity of QKV representations: remains as high as 0.9No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy