Natural Language Inference On Rte

评估指标

Accuracy

评测结果

各个模型在此基准测试上的表现结果

模型名称	Accuracy	Paper Title	Repository
LTG-BERT-small 24M	53.7	Not all layers are equally as important: Every Layer Counts BERT	-
DistilBERT 66M	62.9%	DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
Flipped-3B	71.05	Guess the Instruction! Flipped Learning Makes Language Models Stronger Zero-Shot Learners
ERNIE	68.8%	ERNIE: Enhanced Language Representation with Informative Entities
T5-XXL 11B	92.5%	SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization
ALBERT	89.2%	ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
GPT-NeoX 20B (1-shot)	53.8%	BloombergGPT: A Large Language Model for Finance	-
PaLM 540B (1-shot)	78.7%	PaLM: Scaling Language Modeling with Pathways
KiC-770M	74.00	Knowledge-in-Context: Towards Knowledgeable Semi-Parametric Language Models	-
data2vec	69.9%	data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language
LaMini-F-T5 783M	65%	LaMini-LM: A Diverse Herd of Distilled Models from Large-Scale Instructions
PaLM 540B (0-shot)	72.9%	PaLM: Scaling Language Modeling with Pathways
Q-BERT (Shen et al., 2020)	84.7	Q-BERT: Hessian Based Ultra Low Precision Quantization of BERT	-
PaLM 2-L (1-shot)	79.3%	PaLM 2 Technical Report
RoE-3B	64.01	Exploring the Benefits of Training Expert Language Models over Instruction Tuning
SMART-BERT	71.2%	SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization
RealFormer	73.7%	RealFormer: Transformer Likes Residual Attention
PaLM 2-S (1-shot)	78.7%	PaLM 2 Technical Report
XLNet (single model)	85.9%	XLNet: Generalized Autoregressive Pretraining for Language Understanding
BigBird	75.0%	Big Bird: Transformers for Longer Sequences

0 of 90 row(s) selected.