List your company
ResearchFormerly WizWand

Semantic Textual Similarity on MultiHate (MH)

26.29STS ScoreLoRA
3.03569.072815.1121.1472May 27, 2026
Updated 5mo ago

Evaluation results

MethodLinks
LoRABackbone=LLaMA-3.2-3B, Evaluation Protocol=Fine-tuning2026.0526.29
Prompt-TuningBackbone=Mistral-7B, Evaluation Protocol=Frozen Backbone Methods2026.0521.49
Pretend-CoTBackbone=Mistral-7B, Evaluation Protocol=Training-free Prompting2026.0519.26
LoRABackbone=Qwen2.5-7B, Evaluation Protocol=Fine-tuning2026.0518.84
Prompt-TuningBackbone=Qwen2.5-7B, Evaluation Protocol=Frozen Backbone Methods2026.0518.26
LoRABackbone=LLaMA-3.2-1B, Evaluation Protocol=Fine-tuning2026.0516.76
Prompt-TuningBackbone=LLaMA-3.2-3B, Evaluation Protocol=Frozen Backbone Methods2026.0516.7
PromptEmbedderBackbone=Qwen2.5-7B, Evaluation Protocol=Frozen Backbone Methods2026.0515.46
Pretend-CoTBackbone=LLaMA-3.2-3B, Evaluation Protocol=Training-free Prompting2026.0515.41
LoRABackbone=Mistral-7B, Evaluation Protocol=Fine-tuning2026.0515.25
Pretend-CoTBackbone=Qwen2.5-7B, Evaluation Protocol=Training-free Prompting2026.0514.63
PromptEOLBackbone=Mistral-7B, Evaluation Protocol=Training-free Prompting2026.0514.07
PromptEmbedderBackbone=LLaMA-3.2-1B, Evaluation Protocol=Frozen Backbone Methods2026.0513.3
PromptEOLBackbone=LLaMA-3.2-3B, Evaluation Protocol=Training-free Prompting2026.0511.21
PromptEOLBackbone=Qwen2.5-7B, Evaluation Protocol=Training-free Prompting2026.0511.09
Pretend-CoTBackbone=LLaMA-3.2-1B, Evaluation Protocol=Training-free Prompting2026.0510.71
PromptEOLBackbone=LLaMA-3.2-1B, Evaluation Protocol=Training-free Prompting2026.059.94
PromptEmbedderBackbone=LLaMA-3.2-3B, Evaluation Protocol=Frozen Backbone Methods2026.058.86
PromptEmbedderBackbone=Mistral-7B, Evaluation Protocol=Frozen Backbone Methods2026.057.4
EchoBackbone=LLaMA-3.2-3B, Evaluation Protocol=Training-free Prompting2026.056.63
EchoBackbone=LLaMA-3.2-1B, Evaluation Protocol=Training-free Prompting2026.056.14
EchoBackbone=Mistral-7B, Evaluation Protocol=Training-free Prompting2026.056
EchoBackbone=Qwen2.5-7B, Evaluation Protocol=Training-free Prompting2026.055.32
Prompt-TuningBackbone=LLaMA-3.2-1B, Evaluation Protocol=Frozen Backbone Methods2026.053.93