Semantic Textual Similarity On Sick

评估指标

Spearman Correlation

评测结果

各个模型在此基准测试上的表现结果

模型名称	Spearman Correlation	Paper Title	Repository
Trans-Encoder-BERT-large-bi (unsup.)	0.7133	Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations	-
Mirror-RoBERTa-base (unsup.)	0.706	Fast, Effective, and Self-Supervised: Transforming Masked Language Models into Universal Lexical and Sentence Encoders	-
SRoBERTa-NLI-large	0.7429	Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks	-
Trans-Encoder-RoBERTa-large-cross (unsup.)	0.7163	Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations	-
SimCSE-RoBERTalarge	0.8195	SimCSE: Simple Contrastive Learning of Sentence Embeddings	-
SRoBERTa-NLI-base	0.7446	Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks	-
Dino (STSb/̄	0.6809	Generating Datasets with Pretrained Language Models	-
PromCSE-RoBERTa-large (0.355B)	0.8243	Improved Universal Sentence Embeddings with Prompt-based Contrastive Learning and Energy-based Learning	-
Trans-Encoder-BERT-base-cross (unsup.)	0.6952	Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations	-
Rematch	0.6772	Rematch: Robust and Efficient Matching of Local Knowledge Graphs to Improve Structural and Semantic Similarity	-
SBERT-NLI-base	0.7291	Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks	-
SBERT-NLI-large	0.7375	Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks	-
IS-BERT-NLI	0.6425	An Unsupervised Sentence Embedding Method by Mutual Information Maximization	-
Dino (STS/̄	0.7426	Generating Datasets with Pretrained Language Models	-
PromptEOL+CSE+OPT-13B	0.8206	Scaling Sentence Embeddings with Large Language Models	-
PromptEOL+CSE+OPT-2.7B	0.8129	Scaling Sentence Embeddings with Large Language Models	-
Mirror-BERT-base (unsup.)	0.703	Fast, Effective, and Self-Supervised: Transforming Masked Language Models into Universal Lexical and Sentence Encoders	-
BERTbase-flow (NLI)	0.6544	On the Sentence Embeddings from Pre-trained Language Models	-
SentenceBERT	0.7462	Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks	-
Trans-Encoder-BERT-base-bi (unsup.)	0.7276	Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations	-

0 of 22 row(s) selected.