List your company
ResearchFormerly WizWand

Representational Similarity on STSB 20% (val)

0.558MOT MetricMOT
0.3760.423250.47050.51775Oct 2, 2025
Updated 6mo ago

Evaluation results

MethodLinks
MOTModel 1=Llama-3.2 1B, Model 2=Llama-3.2 3B2025.100.558———
MOTModel 1=Qwen-2.5 0.5B, Model 2=Llama-3.2 3B2025.100.531———
MOTModel 1=Qwen-2.5 0.5B, Model 2=Llama-3.2 1B2025.100.522———
MOTModel 1=Qwen-2.5 0.5B, Model 2=Qwen-2.5 3B2025.100.51———
MOTModel 1=Llama-3.2 1B, Model 2=Qwen-2.5 3B2025.100.432———
MOTModel 1=Llama-3.2 3B, Model 2=Qwen-2.5 3B2025.100.383———
Pairwise Best OTModel 1=Llama-3.2 1B, Model 2=Llama-3.2 3B2025.10———0.505
Pairwise Best OTModel 1=Qwen-2.5 0.5B, Model 2=Qwen-2.5 3B2025.10———0.477
Pairwise Best OTModel 1=Qwen-2.5 0.5B, Model 2=Llama-3.2 1B2025.10———0.511
Pairwise Best OTModel 1=Qwen-2.5 0.5B, Model 2=Llama-3.2 3B2025.10———0.524
Pairwise Best OTModel 1=Llama-3.2 1B, Model 2=Qwen-2.5 3B2025.10———0.38
Pairwise Best OTModel 1=Llama-3.2 3B, Model 2=Qwen-2.5 3B2025.10———0.346
Random (Perm-P)Model 1=Llama-3.2 1B, Model 2=Llama-3.2 3B2025.10—0.51——
Random (Perm-P)Model 1=Qwen-2.5 0.5B, Model 2=Qwen-2.5 3B2025.10—0.494——
Random (Perm-P)Model 1=Qwen-2.5 0.5B, Model 2=Llama-3.2 1B2025.10—0.5——
Random (Perm-P)Model 1=Qwen-2.5 0.5B, Model 2=Llama-3.2 3B2025.10—0.513——
Random (Perm-P)Model 1=Llama-3.2 1B, Model 2=Qwen-2.5 3B2025.10—0.411——
Random (Perm-P)Model 1=Llama-3.2 3B, Model 2=Qwen-2.5 3B2025.10—0.374——
Single-Best OTModel 1=Llama-3.2 1B, Model 2=Llama-3.2 3B2025.10——0.502—
Single-Best OTModel 1=Qwen-2.5 0.5B, Model 2=Qwen-2.5 3B2025.10——0.467—
Single-Best OTModel 1=Qwen-2.5 0.5B, Model 2=Llama-3.2 1B2025.10——0.502—
Single-Best OTModel 1=Qwen-2.5 0.5B, Model 2=Llama-3.2 3B2025.10——0.498—
Single-Best OTModel 1=Llama-3.2 1B, Model 2=Qwen-2.5 3B2025.10——0.345—
Single-Best OTModel 1=Llama-3.2 3B, Model 2=Qwen-2.5 3B2025.10——0.338—