AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

TokenUnify: Scaling Up Autoregressive Pretraining for Neuron Segmentation

Yinda Chen, Haoyuan Shi, Xiaoyu Liu, Te Shi, Ruobing Zhang, Dong Liu, Zhiwei Xiong, Feng Wu

ICCV 2025 · Vision-language and multimodal › Human and embodied modeling · approach: Medical vision-language models · 32 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 next-token prediction to random-token pretraining (ablation)TokenUnify · Wafer4 (Waterz)VOI: 1.2680 -> 1.1300Improved
add next-all token prediction to random-token pretraining (ablation)TokenUnify · Wafer4 (Waterz)VOI: 1.2680 -> 1.1907Improved
add next-all token prediction to random + next-token pretraining (ablation)TokenUnify · Wafer4 (Waterz)VOI: 1.1300 -> 0.9951Improved
add next-all token prediction to random + next-token pretraining (ablation)TokenUnify · Wafer4 (Waterz)ARAND: 0.0692 -> 0.0509Improved
use next-token prediction alone instead of random-token prediction alone (ablation)TokenUnify · Wafer4 (Waterz)VOI: 1.2680 -> 4.0418Hurt
add decoder to fine-tuning of Mamba + encoder (ablation)TokenUnify · Wafer4 (Waterz)VOI: 1.1065 -> 0.9951Improved
fine-tune Mamba blocks only rather than encoder only (ablation)TokenUnify · Wafer4 (Waterz)VOI: 1.5556 -> 1.1362Improved
fine-tune Mamba blocks only rather than decoder only (ablation)TokenUnify · Wafer4 (Waterz)VOI: 1.5295 -> 1.1362Improved
use TokenUnify pretraining instead of no pretrainingEMmamba · Wafer4 (Waterz)VOI: 1.7839 -> 0.9951Improved
use TokenUnify pretraining instead of MAE pretrainingEMmamba · Wafer4 (Waterz)VOI: 1.3248 -> 0.9951Improved
use TokenUnify pretraining instead of MAE pretrainingEMmamba · Wafer4 (LMC)VOI: 2.0709 -> 1.6521Improved
use TokenUnify pretraining instead of MAE pretrainingEMmamba · AC3/4 (Waterz)VOI: 1.1131 -> 0.9918Improved

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