LEGO (Parameter-efficient fine-tuning (LoRA family)): current frontier — recent, not yet superseded in the knowledge base. 0 paper(s) critique it, 0 beat it on benchmarks — not ranked as a superseded baseline. Sub-problem: cluster led by CNN. Newer alternatives in the same sub-problem include Structured Convolutional Projection + LoRA, LEGO, layer-selective multimodal large language models (MLLMs) with contrastive LoRA tuning and layer sensitivity analysis (LSA), AdaptPrompt.

Parameter-efficient fine-tuning (LoRA family)

LEGO

LEGO: LoRA-Enabled Generator-Oriented Framework for Synthetic Image Detection

On the frontierfirst seen May 6, 2026

Current frontier — recent, not yet superseded in the knowledge base

0 papers critique it · 0 beat it on benchmarks

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.

Structured Convolutional Projection + LoRAEfficient and Adaptive Human Activity Recognition via LLM Backbones
May 12, 2026
LEGOLEGO: LoRA-Enabled Generator-Oriented Framework for Synthetic Image Detection
May 6, 2026
layer-selective multimodal large language models (MLLMs) with contrastive LoRA tuning and layer sensitivity analysis (LSA)Fine-Grained Human Pose Editing Assessment via Layer-Selective MLLMs
Jan 15, 2026
AdaptPromptAdaptPrompt: Parameter-Efficient Adaptation of VLMs for Generalizable Deepfake Detection
Dec 19, 2025