List your company
ResearchFormerly WizWand

Natural Language Inference on RTE (val)

0.918AccuracyUnified PPT
0.473920.589210.70450.81979Sep 9, 2021Feb 24, 2022Aug 12, 2022Jan 28, 2023Jul 15, 2023Dec 31, 2023Jun 17, 2024
Updated 8mo ago

Evaluation results

MethodLinks
Unified PPTtuning_strategy=unified pre-trained prompt tuning2021.090.918——
FTtuning_strategy=fine-tuning2021.090.898——
PPTtuning_strategy=pre-trained prompt tuning2021.090.896——
PTtuning_strategy=prompt-tuning2021.090.88——
ENSBackbone=RoBERTa-base2024.060.77740.19471.04
MCDBackbone=RoBERTa-base2024.060.76650.13060.65
LAPBackbone=RoBERTa-base2024.060.76050.05330.55
BLoBBackbone=RoBERTa-base, N=02024.060.76050.14640.56
MLEBackbone=RoBERTa-base2024.060.75810.20591.11
MAPBackbone=RoBERTa-base2024.060.75810.21671.23
BLoBBackbone=RoBERTa-base, N=102024.060.75450.08970.48
BLoBBackbone=RoBERTa-base, N=52024.060.74610.10460.5
ZeroGenPLM=GPT2-XL, TAM=DistilBERT, Setting=ZEROGEN, #Param=66M2022.020.5993——
ZeroGenPLM=GPT2-XL, TAM=LSTM, Setting=ZEROGEN, #Param=~7M2022.020.5885——
ZeroGenPLM=GPT2-Large, TAM=DistilBERT, Setting=ZEROGEN, #Param=66M2022.020.5848——
DistilBERTPLM=Gold Data, Setting=SUPERVISED, #Param=66M2022.020.5812——
GPT2-XLSetting=PROMPTING, #Param=1.5B2022.020.5704——
ZeroGenPLM=GPT2-Large, TAM=LSTM, Setting=ZEROGEN, #Param=~7M2022.020.5668——
LSTMPLM=Gold Data, Setting=SUPERVISED, #Param=~7M2022.020.5487——
GPT2-LargeSetting=PROMPTING, #Param=762M2022.020.5451——
GPT2Setting=PROMPTING, #Param=117M2022.020.527——
ZeroGenPLM=GPT2, TAM=DistilBERT, Setting=ZEROGEN, #Param=66M2022.020.5054——
BBBBackbone=RoBERTa-base2024.060.49460.02660.69
ZeroGenPLM=GPT2, TAM=LSTM, Setting=ZEROGEN, #Param=~7M2022.020.491——