List your company
ResearchFormerly WizWand

Word Similarity on RW-STANFORD

0.4Spearman CorrelationW2G
0.152480.216740.2810.34526Oct 17, 2018Jan 25, 2020May 4, 2021Aug 12, 2022Nov 20, 2023Feb 27, 2025Jun 7, 2026
Updated 4mo ago

Evaluation results

MethodLinks
W2GEmbedding Method=W2G, Training Corpus=ukWaC and WaCkypedia2018.100.4
GloVe+GTEmbedding Method=GloVe, Transform=Gaussian Transform, Refinement Corpus=text82018.100.38
GloVeEmbedding Method=GloVe2018.100.34
EllEmbedding Method=Ell, Training Corpus=ukWaC and WaCkypedia2018.100.29
HALO-CBOWembedding dimension=200, core vocabulary size=20000, Trainable Parameters=33,660,3222026.060.253
HALO-SGNSembedding dimension=200, core vocabulary size=20000, Trainable Parameters=33,660,3222026.060.253
W2V*text8Embedding Method=W2V, Training Corpus=text82018.100.25
HALO-CBOWembedding dimension=200, core vocabulary size=10000, Trainable Parameters=29,660,3222026.060.243
HALO-CBOWembedding dimension=200, core vocabulary size=30000, Trainable Parameters=37,660,3222026.060.233
CBOWembedding dimension=200, Trainable Parameters=200,000,0002026.060.203
GloVe*text8Embedding Method=GloVe, Training Corpus=text82018.100.2
CBOWembedding dimension=100, Trainable Parameters=100,000,0002026.060.192
CBOWembedding dimension=50, Trainable Parameters=50,000,0002026.060.162