List your company
ResearchFormerly WizWand

Reading Comprehension on RACE mid

89.9AccuracyFlexora
73.457677.726381.99586.2637Mar 29, 2022Nov 14, 2022Jul 3, 2023Feb 19, 2024Oct 7, 2024May 26, 2025Jan 12, 2026
Updated 8mo ago

Evaluation results

MethodLinks
FlexoraBackbone=Qwen1.5-7B2024.0889.9
FlexoraBackbone=Gemma-7B2024.0889.62
FlexoraBackbone=Yi-6B2024.0888.72
Flexora (w/ LoReFT)Backbone Model=Llama3-8B, LoRA Algorithm=LoReFT, Flexora Integration=true2024.0888.36
Flexora (w/ rsLoRA)Backbone Model=Llama3-8B, LoRA Algorithm=rsLoRA, Flexora Integration=true2024.0888.21
LoReFTBackbone Model=Llama3-8B, LoRA Algorithm=LoReFT, Flexora Integration=false2024.0888.21
KALEBackbone=Gemma2 9B, Category=Augmented-based2026.0187.74
rsLoRABackbone Model=Llama3-8B, LoRA Algorithm=rsLoRA, Flexora Integration=false2024.0887.6
LoRABackbone=Qwen1.5-7B2024.0887.59
FlexoraBackbone=Qwen2.5-32B2024.0887.16
Flexora (w/ DoRA)Backbone Model=Llama3-8B, LoRA Algorithm=DoRA, Flexora Integration=true2024.0887.12
Chinchillaevaluation_mode=Few-Shot2022.0386.8
DoRABackbone Model=Llama3-8B, LoRA Algorithm=DoRA, Flexora Integration=false2024.0886.77
KALEBackbone=Qwen2.5 32B, Category=Augmented-based2026.0186.7
FlexoraBackbone=Zephyr-7B-beta2024.0886.19
MoSLoRABackbone Model=Llama3-8B, LoRA Algorithm=MoSLoRA, Flexora Integration=false2024.0886.13
LoRABackbone Model=Llama3-8B, LoRA Algorithm=LoRA, Flexora Integration=false2024.0885.86
LoRABackbone=Gemma-7B2024.0885.73
FlexoraBackbone=Mistral-7B-v0.12024.0885.72
FlexoraLoRA Rank (r)=162024.0885.62
Flexora (w/ MoSLoRA)Backbone Model=Llama3-8B, LoRA Algorithm=MoSLoRA, Flexora Integration=true2024.0885.09
FlexoraBackbone=Llama2-7B2024.0884.89
FlexoraBackbone=Xuan Yuan-6B2024.0884.89
Flexora (w/ LoRA)Backbone Model=Llama3-8B, LoRA Algorithm=LoRA, Flexora Integration=true2024.0884.61
FlexoraBackbone=Llama3-8B2024.0884.61
FlexoraLoRA Rank (r)=82024.0884.61
FlexoraFine-tuning Strategy=LoRA, Enhanced=Flexora2024.0884.61
Flexora(w/ Base Model)Backbone=Meta-Llama-3-8B, Model variant=Base, Adaptation=Flexora2024.0884.61
Flexora(w/ Instruct Model)Backbone=Meta-Llama-3-8B-Instruct, Model variant=Instruct, Adaptation=Flexora2024.0884.58
FlexoraLoRA Rank (r)=322024.0884.27
LoRABackbone=Qwen2.5-32B2024.0884.17
GPT3MixBackbone=Gemma2 9B, Category=Augmented-based2026.0183.98
LoRABackbone=Mistral-7B-v0.12024.0883.79
Random (Greedy)Selection=randomly selected layers, LoRA Rank (r)=82024.0883.77
VeRALoRA Rank (r)=82024.0883.55
LoRABackbone=Zephyr-7B-beta2024.0883.45
AugGPTBackbone=Gemma2 9B, Category=Augmented-based2026.0183.33
LoRA-SPLoRA Rank (r)=82024.0883.27
AdaLoRAro=4, LoRA Rank (r)=82024.0883.15
FlexoraFine-tuning Strategy=Full Fine-Tuning, Enhanced=Flexora2024.0883.13
DMTBackbone=Gemma2 9B, Category=SFT-based2026.0183.01
LoRAPruneRatio=0.5, LoRA Rank (r)=82024.0882.96
FlexoraBackbone=Vicuna-7B-v1.52024.0882.94
LoRABackbone=Llama3-8B2024.0882.86
LoRALoRA Rank (r)=82024.0882.86
LoRAFine-tuning Strategy=LoRA2024.0882.86
LoRA-FALoRA Rank (r)=82024.0882.79
KG-SFTBackbone=Gemma2 9B, Category=SFT-based2026.0182.73
LoRALoRA Rank (r)=322024.0882.36
Full FTFine-tuning protocol=full parameter fine-tuning2024.0881.92
Full FTFine-tuning Strategy=Full Fine-Tuning2024.0881.92
AugGPTBackbone=Qwen2.5 32B, Category=Augmented-based2026.0181.89
MeanLearnBackbone=Qwen2.5 32B, Category=SFT-based2026.0181.82
GraphRAGBackbone=Gemma2 9B, Category=Retrieval-based2026.0181.69
LoRALoRA Rank (r)=162024.0881.63
FlexoraBackbone=Llama-7B2024.0881.62
STaRBackbone=Gemma2 9B, Category=Augmented-based2026.0181.55
Instruct ModelBackbone=Meta-Llama-3-8B-Instruct, Model variant=Instruct2024.0881.36
StructGPTBackbone=Gemma2 9B, Category=Retrieval-based2026.0181.27
SDFTBackbone=Gemma2 9B, Category=SFT-based2026.0181.2
MeanLearnBackbone=Gemma2 9B, Category=SFT-based2026.0181.2
SFTBackbone=Gemma2 9B, Category=SFT-based2026.0180.78
GPT3MixBackbone=Qwen2.5 32B, Category=Augmented-based2026.0180.77
SDFTBackbone=Qwen2.5 32B, Category=SFT-based2026.0180.71
KG-SFTBackbone=Qwen2.5 32B, Category=SFT-based2026.0180.43
STaRBackbone=Qwen2.5 32B, Category=Augmented-based2026.0180.43
LoRABackbone=Yi-6B2024.0880.02
KALEBackbone=Orca2 7B, Category=Augmented-based2026.0180.02
DMTBackbone=Qwen2.5 32B, Category=SFT-based2026.0180.01
TOGBackbone=Gemma2 9B, Category=Retrieval-based2026.0179.53
SFTBackbone=Qwen2.5 32B, Category=SFT-based2026.0179.39
LoRABackbone=Llama2-7B2024.0879.32
CoTBackbone=Gemma2 9B, Category=Prompt-based2026.0179.32
FlexoraBackbone=Chatglm3-6B2024.0879.18
MeanLearnBackbone=Orca2 7B, Category=SFT-based2026.0178.76
VanillaBackbone=Gemma2 9B, Category=Prompt-based2026.0178.62
AugGPTBackbone=Orca2 7B, Category=Augmented-based2026.0178.55
StructGPTBackbone=Orca2 7B, Category=Retrieval-based2026.0178.41
TOGBackbone=Qwen2.5 32B, Category=Retrieval-based2026.0178.34
LoRABackbone=Xuan Yuan-6B2024.0878.04
StructGPTBackbone=Qwen2.5 32B, Category=Retrieval-based2026.0177.92
CoTBackbone=Qwen2.5 32B, Category=Prompt-based2026.0177.64
GPT3MixBackbone=Orca2 7B, Category=Augmented-based2026.0177.51
GraphRAGBackbone=Qwen2.5 32B, Category=Retrieval-based2026.0177.37
LoRA-dropLoRA Rank (r)=82024.0877.3
DMTBackbone=Orca2 7B, Category=SFT-based2026.0177.3
STaRBackbone=Orca2 7B, Category=Augmented-based2026.0177.3
VanillaBackbone=Qwen2.5 32B, Category=Prompt-based2026.0176.95
KG-SFTBackbone=Orca2 7B, Category=SFT-based2026.0176.9
SFTBackbone=Orca2 7B, Category=SFT-based2026.0176.88
KALEBackbone=OLMOE 7B, Category=Augmented-based2026.0175.84
CoTBackbone=Orca2 7B, Category=Prompt-based2026.0175.84
GraphRAGBackbone=Orca2 7B, Category=Retrieval-based2026.0175.77
LoRABackbone=Llama-7B2024.0875.69
VanillaBackbone=Orca2 7B, Category=Prompt-based2026.0175.49
GPT3MixBackbone=OLMOE 7B, Category=Augmented-based2026.0175.48
SDFTBackbone=Orca2 7B, Category=SFT-based2026.0175.14
Gopherevaluation_mode=Few-Shot2022.0375.1
KALEBackbone=LlaMA3 8B, Category=Augmented-based2026.0174.12
TOGBackbone=Orca2 7B, Category=Retrieval-based2026.0174.09