Method Drift›Mixture-of-experts routing
Superseded baseline#150 of 1,370 most-superseded
Loss-free balancing
Mixture-of-experts routing
superseded — cited as a baseline and beaten by newer methods
1 papers critique it · 1 beat it on benchmarks
What papers say
Verbatim critique sentences, each from a paper that cites Loss-free balancing as a baseline.
“Loss-free routing exhibits the largest imbalance in activation frequencies, with a few experts over-selected and many under-used, indicating the strongest capacity imbalance and likely explaining its weaker accuracy”
— Hierarchical Mixture-of-Experts with Two-Stage Optimization
Beaten on benchmarks
Head-to-head results where a newer method reports beating Loss-free balancing. Values are copied from the source paper's tables — verify against the cited paper.
- Hierarchical Mixture-of-Experts with Two-Stage Optimization
Hi-MoE beats Loss-free balancing · Acc@1 [Swin-MoE on Tiny ImageNet]
59.13 vs 58.87
- Hierarchical Mixture-of-Experts with Two-Stage Optimization
Hi-MoE beats Loss-free balancing · PPL [nanoGPT on OpenWebText]
2.947 vs 2.979
Newer alternatives
Recent methods in the same sub-problem, not yet superseded in the knowledge base.