List your company

Multitask Evaluation on Aggregate (MMLU, GSM8K, SQuAD V2, Amazon, 20News)

77.11Average ScoreShadowPEFT w/ pretrained shadow
33.4344.7756.1167.45Apr 21, 2026

Evaluation results

MethodLinks
ShadowPEFT w/ pretrained shadowTrainable Params=454,991,872, Backbone Model=Qwen3 8B + 0.5B Shadow, Inference Mode=Shadow-attached, Shadow Model Initialization=pretrained2026.0477.11
ShadowPEFTTrainable Params=29,118,720, Backbone Model=Qwen3 8B, Inference Mode=Shadow-attached2026.0476.92
ShadowPEFT w/ random shadowTrainable Params=454,991,872, Backbone Model=Qwen3 8B + 0.5B Shadow, Inference Mode=Shadow-attached, Shadow Model Initialization=random2026.0476.6
LoRATrainable Params=30,670,848, Backbone Model=Qwen3 8B2026.0476.51
DoRATrainable Params=31,039,488, Backbone Model=Qwen3 8B2026.0475.99
ShadowPEFTTrainable Params=23,442,176, Backbone Model=Qwen3 4B, Inference Mode=Shadow-attached2026.0475.43
DoRATrainable Params=23,906,304, Backbone Model=Qwen3 4B2026.0474.85
LoRATrainable Params=23,592,960, Backbone Model=Qwen3 4B2026.0474.55
ShadowPEFTTrainable Params=9,073,920, Backbone Model=Qwen3 0.6B, Inference Mode=Shadow-attached2026.0462.27
Detached Shadow Only (pretrained)Trainable Params=454,991,872, Backbone Model=Qwen3 8B + 0.5B Shadow, Inference Mode=Detached, Shadow Model Initialization=pretrained2026.0462.11
DoRATrainable Params=9,318,400, Backbone Model=Qwen3 0.6B2026.0462.08
LoRATrainable Params=9,175,040, Backbone Model=Qwen3 0.6B2026.0461.81
Detached Shadow OnlyTrainable Params=23,442,176, Backbone Model=Qwen3 4B, Inference Mode=Detached2026.0438.4
Detached Shadow OnlyTrainable Params=9,073,920, Backbone Model=Qwen3 0.6B, Inference Mode=Detached2026.0436.79
Detached Shadow OnlyTrainable Params=29,118,720, Backbone Model=Qwen3 8B, Inference Mode=Detached2026.0436.09
Detached Shadow Only (random)Trainable Params=454,991,872, Backbone Model=Qwen3 8B + 0.5B Shadow, Inference Mode=Detached, Shadow Model Initialization=random2026.0435.11