Pith. sign in

REVIEW 10 cited by

Upcycling Large Language Models into Mixture of Experts

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.07524 v2 pith:LUU4NGDJ submitted 2024-10-10 cs.CL cs.AIcs.LG

Upcycling Large Language Models into Mixture of Experts

classification cs.CL cs.AIcs.LG
keywords upcyclingmodelsmodellanguageapproachtrainedachieveddense
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Upcycling pre-trained dense language models into sparse mixture-of-experts (MoE) models is an efficient approach to increase the model capacity of already trained models. However, optimal techniques for upcycling at scale remain unclear. In this work, we conduct an extensive study of upcycling methods and hyperparameters for billion-parameter scale language models. We propose a novel "virtual group" initialization scheme and weight scaling approach to enable upcycling into fine-grained MoE architectures. Through ablations, we find that upcycling outperforms continued dense model training. In addition, we show that softmax-then-topK expert routing improves over topK-then-softmax approach and higher granularity MoEs can help improve accuracy. Finally, we upcycled Nemotron-4 15B on 1T tokens and compared it to a continuously trained version of the same model on the same 1T tokens: the continuous trained model achieved 65.3% MMLU, whereas the upcycled model achieved 67.6%. Our results offer insights and best practices to effectively leverage upcycling for building MoE language models. Code is available.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 10 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Expert Upcycling: Shifting the Compute-Efficient Frontier of Mixture-of-Experts

    cs.LG 2026-04 unverdicted novelty 7.0

    Expert upcycling duplicates experts in an existing MoE checkpoint and continues pre-training to match fixed-size baseline performance with 32% less compute.

  2. Expert Upcycling: Shifting the Compute-Efficient Frontier of Mixture-of-Experts

    cs.LG 2026-04 unverdicted novelty 6.0

    Expert upcycling expands MoE models by duplicating experts and continuing pre-training, matching baseline performance while saving 32% GPU hours in 7B-13B experiments.

  3. SpikingBrain: Spiking Brain-inspired Large Models

    cs.LG 2025-09 unverdicted novelty 6.0

    SpikingBrain-7B and SpikingBrain-76B achieve Transformer-comparable performance after continual pre-training on 150B tokens, with over 100x TTFT speedup on 4M-token sequences and 69.15% sparsity from event-driven spiking.

  4. Mixture-of-Experts Can Surpass Dense LLMs Under Strictly Equal Resource

    cs.CL 2025-06 conditional novelty 6.0

    MoE models with activation rates in an optimal region outperform dense LLMs of identical total parameter count, training compute, and data budget, with the optimal region consistent across scales.

  5. MoLGE: Mixture of Language Group Experts for Efficient Scaling of Massively Multilingual Speech Recognition

    cs.CL 2026-07 conditional novelty 5.0

    Grouping 495 languages into roughly 16 clusters and routing speech to group-specific LoRA experts improves multilingual ASR error rates over dense and random baselines.

  6. Tackling Multimodal Learning Challenges with Mixture-of-Expert: A Survey

    cs.LG 2026-05 accept novelty 5.0

    A literature survey that categorizes how Mixture-of-Experts architectures address multimodal learning challenges and identifies open research gaps.

  7. Beyond Sunk Costs: Boosting LLM Pre-training Efficiency via Orthogonal Growth of Mixture-of-Experts

    cs.LG 2025-10 unverdicted novelty 5.0

    Orthogonal growth recycles pre-trained MoE checkpoints via layer copying and noisy expert duplication, delivering 10.6% higher accuracy than training from scratch with equivalent extra compute.

  8. Reversible Foundations: Training a 120B Sparse MoE through State-Preserving Scaling

    cs.LG 2026-06 unverdicted novelty 4.0

    A 120B sparse MoE model with 460 experts was trained on one 8-GPU node to loss 1.78 using reversible recurrence and state-preserving scaling from a 1.78B dense seed, with 5.93B active parameters.

  9. ScatterPrism: convergence for generative simulation and inverse problems in particle and nuclear physics

    cs.LG 2026-04 unverdicted novelty 4.0

    CFM training loss plateaus while physics-informed metrics continue improving; ScatterPrism and a multi-metric protocol are proposed to restore kinematic fidelity without memorization.

  10. ScatterPrism: convergence for generative simulation and inverse problems in particle and nuclear physics

    cs.LG 2026-04 unverdicted novelty 4.0

    Standard Conditional Flow Matching loss is a misleading early plateau; physics-informed metrics keep improving, so ScatterPrism and multi-metric diagnostics are needed for kinematic fidelity.