List your company
ResearchFormerly WizWand

Commonsense Reasoning on aNLI

87.3AccuracyRAINBOW
30.630445.342760.05574.7673Oct 17, 2022May 12, 2023Dec 6, 2023Jul 1, 2024Jan 25, 2025Aug 21, 2025Mar 17, 2026
Updated 7mo ago

Evaluation results

MethodLinks
RAINBOWknowledge-augmented=true, plausibility annotations=with2023.0587.3
RoBERTaplausibility annotations=with2023.0585.6
DRAGON2022.1084
GreaseLM2022.1083.3
QAGNN2022.1083
RoBERTa2022.1082.7
LiPoRplausibility annotations=without2023.0571.56
KDDC-ATOMIC (N)knowledge-augmented=true, plausibility annotations=without2023.0570.8
KDDC-CSKG (N)knowledge-augmented=true, plausibility annotations=without2023.0570.5
KDDC-CWWV (N)knowledge-augmented=true, plausibility annotations=without2023.0570
ZS GPT3setting=zero-shot, plausibility annotations=without2023.0567.54
Previous Bestplausibility annotations=without2023.0565.5
CKT-baseBackbone=T5-base, Method=CKT2023.0664.37
CALMBackbone=T5-base2023.0663.2
KnowBERTBackbone=KnowBERT2023.0663.18
ERNIE-baseBackbone=ERNIE-base2023.0663.04
BERT-baseBackbone=BERT-base2023.0661.88
T5-base + SSMBackbone=T5-base, Augmentation=SSM2023.0661.57
T5-base + TIBackbone=T5-base, Augmentation=TI2023.0661.42
T5-baseBackbone=T5-base2023.0661.1
T5-base + KDBackbone=T5-base, Augmentation=KD2023.0660.94
T5-base + CSKG (Rule)Backbone=T5-base, Augmentation=CSKG (Rule)2023.0660.15
CKT w/ GPT-2Backbone=GPT-2, Method=CKT2023.0660.14
T5-base + CSKG (TI)Backbone=T5-base, Augmentation=CSKG (TI)2023.0659.92
COMETBackbone=BART2023.0657.63
ZS GPT-NEOsetting=zero-shot, plausibility annotations=without2023.0557.47
Tuned BARTsetting=tuned, plausibility annotations=without2023.0557.4
ZS BARTsetting=zero-shot, plausibility annotations=without2023.0550.96
MDM-Prime-v2Zero-shot=true, Number of Parameters=1.1B2026.0334.24
OPTZero-shot=true, Number of Parameters=1.3B2026.0333.86
BloomZero-shot=true, Number of Parameters=1.1B2026.0333.35
PythiaZero-shot=true, Number of Parameters=1.4B2026.0333.29
TinyLLaMAZero-shot=true, Number of Parameters=1.1B2026.0333.25
GPT-NeoZero-shot=true, Number of Parameters=1.3B2026.0333.13
SMDMZero-shot=true, Number of Parameters=1.1B2026.0332.81