AI Paper Atlas

Research map / Architectures and efficiency / Mixture of experts

DND: Boosting Large Language Models with Dynamic Nested Depth

Tieyuan Chen, Xiaodong Chen, Haoxing Chen, Zhenzhong Lan, Weiyao Lin, Jianguo Li

ICLR 2026 · Architectures and efficiency › Mixture of experts · approach: Efficient MoE language model systems

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 DND post-training to SFTQwen3-1.7B · Dense SFT average (11 benchmarks)average score: 59.53 -> 61.41Improved
add DND post-training to SFTLlama3.2-1B · Dense SFT average (11 benchmarks)average score: 45.37 -> 47.98Improved
add DND post-training to SFTGemma3-1B · Dense SFT average (11 benchmarks)average score: 47.08 -> 49.58Improved
add DND post-training to SFTQwen3-30B-A3B · MoE SFT average (17 benchmarks)average score: 75.70 -> 76.57Improved
use ITT token selection instead of DND under same computeQwen3-1.7B · Dense SFT average (11 benchmarks)average score: 59.53 -> 59.58No clear effect
add router control loss and threshold control to DND (ablation)DND on Qwen3-1.7B · Dense SFT average (11 benchmarks)average score: 60.54 -> 61.41Improved
add threshold control on top of router control (ablation)DND on Qwen3-1.7B · Dense SFT average (11 benchmarks)average score: 60.68 -> 61.41Improved
add router control loss on top of threshold control (ablation)DND on Qwen3-1.7B · Dense SFT average (11 benchmarks)average score: 60.58 -> 61.41Improved
lower target selection ratio from 20% to 10% (ablation)DND on Qwen3-1.7B · Dense SFT average (11 benchmarks)average score: 61.41 -> 60.33Hurt
raise target selection ratio from 20% to 30% (ablation)DND on Qwen3-1.7B · Dense SFT average (11 benchmarks)average score: 61.41 -> 61.03No clear effect
apply DND to layers 3:24 instead of 4:23 (ablation)DND on Qwen3-1.7B · Dense SFT average (11 benchmarks)average score: 61.41 -> 60.36Hurt
apply DND to layers 5:22 instead of 4:23 (ablation)DND on Qwen3-1.7B · Dense SFT average (11 benchmarks)average score: 61.41 -> 61.05Hurt

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