Method Drift›Parameter-efficient fine-tuning (LoRA family)
MoLoRA
Parameter-efficient fine-tuning (LoRA family)
superseded — cited as a baseline and beaten by newer methods
1 papers critique it · 2 beat it on benchmarks
What papers say
Verbatim critique sentences, each from a paper that cites MoLoRA as a baseline.
“However, MoLoRA introduces additional challenges, such as increased training latency and parameter redundancy.”
— MALoRA: Mixture of Asymmetric Low-Rank Adaptation for Enhanced Multi-Task Learning
Beaten on benchmarks
Head-to-head results where a newer method reports beating MoLoRA. Values are copied from the source paper's tables — verify against the cited paper.
- MALoRA: Mixture of Asymmetric Low-Rank Adaptation for Enhanced Multi-Task Learning
MALoRA beats MoLoRA · AVG [inter-domain multi-task]
56.3 vs 55.3
- MALoRA: Mixture of Asymmetric Low-Rank Adaptation for Enhanced Multi-Task Learning
MALoRA beats MoLoRA · AVG [intra-domain common-sense reasoning]
80.47 vs 79.69
- VLA-GSE: Boosting Parameter-Efficient Fine-Tuning in VLA with Generalized and Specialized Experts
VLA-GSE beats MoLoRA · Total success rate [Comparable parameter budget (~2.5%)]
81.2 vs 76.2
- VLA-GSE: Boosting Parameter-Efficient Fine-Tuning in VLA with Generalized and Specialized Experts
VLA-GSE beats MoLoRA · Average success rate [LIBERO standard evaluation]
98.4 vs 95.8
Newer alternatives
Recent methods in the same sub-problem, not yet superseded in the knowledge base.