Our new X account is live! Follow @wizwand_team for updates

Text Classification on IMDB (test)

96.8CA

Protoformer

Updated 3d ago

Evaluation Results

Method
Protoformer 2022.06	96.8	-	-	-	-	96.4	95.2
RoBERTa 2022.06	95.7	-	-	-	-	95.2	94.1
BERT 2021.05	95.4	-	-	-	-	-	-
Standard 2021.12	94.9	-	-	-	-	-	-
CHARFORMER SBase 2021.06	94.4	-	-	-	-	-	-
T5 Base, Subword 2021.06	94.2	-	-	-	-	-	-
DocBERT 2022.06	93.6	-	-	-	-	93.2	92.1
SAFER 2025.07	93.5	9.5	-	-	-	-	-
PGD-K 2025.07	93.2	26	-	-	-	-	-
RanMASK 2025.07	93.2	22	-	-	-	-	-
FreeLB++ 2025.07	93.2	45.3	-	-	-	-	-
Standard 2021.12	93.1	-	-	-	-	-	-
FreeLB 2025.07	93	29	-	-	-	-	-
TA-VAT 2025.07	93	28	-	-	-	-	-
Attention (SMaT) 2022.04	92.84	-	-	-	-	-	-
Attention (SMaT) 2022.04	92.56	-	-	-	-	-	-
baseline (BERT) 2025.07	92.1	10.3	-	-	-	-	-
GBM 2025.07	92.1	90.6	-	-	-	-	-
InfoBERT 2025.07	92	19	-	-	-	-	-
MixADA 2025.07	91.9	19	-	-	-	-	-
Text-CRS 2025.07	91.5	84.4	-	-	-	-	-
Byte-level T5 Base 2021.06	91.5	-	-	-	-	-	-
CHARFORMER Base 2021.06	91.5	-	-	-	-	-	-
Attention (SMaT) 2022.04	91.48	-	-	-	-	-	-
Attention (last layer) 2022.04	91.47	-	-	-	-	-	-
Drop Clause Tsetlin Machine 2021.05	91.27	-	-	-	-	-	-
Byte-level T5+LASC Base 2021.06	91.1	-	-	-	-	-	-
DNE 2025.07	90.4	28	-	-	-	-	-
GBM 2025.07	90.2	-	76.3	76.6	84.3	-	-
Attention (all layers) 2022.04	90.08	-	-	-	-	-	-
Standard 2025.07	89.7	-	0.6	2.6	1.4	-	-
GBM 2025.07	89.6	-	76.8	77.8	84.3	-	-
HAN 2022.06	89.6	-	-	-	-	89.4	88.2
Standard 2025.07	89.1	-	0.2	1.6	0.3	-	-
BiLSTM 2021.05	88.9	-	-	-	-	-	-
CNN 2021.05	87.9	-	-	-	-	-	-
ASCC 2025.07	87.8	19.4	-	-	-	-	-
SEM 2025.07	87.6	-	62.2	63.5	61.5	-	-
SEM 2025.07	86.8	-	61.9	63.7	62.2	-	-
Attention (all layers) 2022.04	85.72	-	-	-	-	-	-
ATFL 2025.07	85.1	-	72.2	75.5	74	-	-
ATFL 2025.07	85	-	63.6	66.8	64.7	-	-
No Explainer 2022.04	84.84	-	-	-	-	-	-
Gradient x Input 2022.04	84.83	-	-	-	-	-	-
ASCC 2025.07	84.8	-	74	75.5	74.5	-	-
Gradient L2 2022.04	84.78	-	-	-	-	-	-
ASCC 2025.07	84.3	-	74.2	76.8	75.5	-	-
RIFT 2021.12	84.2	-	-	-	-	-	-
Integrated Gradients 2022.04	84.2	-	-	-	-	-	-
Gradient x Input 2022.04	83.84	-	-	-	-	-	-
No Explainer 2022.04	83.44	-	-	-	-	-	-
Attention (last layer) 2022.04	83.15	-	-	-	-	-	-
Attention (all layers) 2022.04	83	-	-	-	-	-	-
Integrated Gradients 2022.04	82.99	-	-	-	-	-	-
Gradient L2 2022.04	82.98	-	-	-	-	-	-
Integrated Gradients 2022.04	81.79	-	-	-	-	-	-
No Explainer 2022.04	81.72	-	-	-	-	-	-
IBP 2025.07	81.7	-	75.9	76	75.9	-	-
Gradient L2 2022.04	81.66	-	-	-	-	-	-
Gradient x Input 2022.04	81.15	-	-	-	-	-	-
Attention (last layer) 2022.04	80.91	-	-	-	-	-	-
Adv-PTWD 2021.12	80.7	-	-	-	-	-	-
Adv-Base 2021.12	80.1	-	-	-	-	-	-
Adv-Mixout 2021.12	79	-	-	-	-	-	-
RIFT 2021.12	78.3	-	-	-	-	-	-
Adv-Mixout 2021.12	77.8	-	-	-	-	-	-
IBP 2025.07	77.6	-	67.5	67.8	67.6	-	-
Adv-PTWD 2021.12	76.6	-	-	-	-	-	-
SVM 2022.06	74.8	-	-	-	-	74.4	73.3
Adv-Base 2021.12	74.6	-	-	-	-	-	-
Standard DR-AGG 2018.06	45.1	-	-	-	-	-	-
Reverse DR-AGG 2018.06	44.5	-	-	-	-	-	-
UPNN(full) 2018.06	43.5	-	-	-	-	-	-
Self-attention 2018.06	43.3	-	-	-	-	-	-
Cached LSTM 2018.06	42.1	-	-	-	-	-	-
Max pooling 2018.06	41.1	-	-	-	-	-	-
UPNN(np UP) 2018.06	40.5	-	-	-	-	-	-
RNTN+Recurrent 2018.06	40	-	-	-	-	-	-
Average pooling 2018.06	39.1	-	-	-	-	-	-
DASH 2023.05	-	-	-	-	-	93.34	-
DASH 2023.05	-	-	-	-	-	93.3	-
FIXMATCH 2023.05	-	-	-	-	-	95.26	-
FIXMATCH 2023.05	-	-	-	-	-	94.28	-
FLEXMATCH 2023.05	-	-	-	-	-	95.22	-
FLEXMATCH 2023.05	-	-	-	-	-	94.84	-
ADAMATCH 2023.05	-	-	-	-	-	95.2	-
ADAMATCH 2023.05	-	-	-	-	-	94.94	-
Prompt-based FT (hard) 2023.05	-	-	-	-	-	86.78	-
Prompt-based FT (hard) 2023.05	-	-	-	-	-	89.52	-
Prompt-based FT (hard) + PCP 2023.05	-	-	-	-	-	92.49	-
Prompt-based FT (hard) + PCP 2023.05	-	-	-	-	-	94.24	-
Prompt-based FT (soft) 2023.05	-	-	-	-	-	88.14	-
Prompt-based FT (soft) 2023.05	-	-	-	-	-	90.8	-
Prompt-based FT (soft) + PCP 2023.05	-	-	-	-	-	93.53	-
Prompt-based FT (soft) + PCP 2023.05	-	-	-	-	-	94.36	-
Prompt-based FT (hard) 2023.05	-	-	-	-	-	95.6	-
Prompt-based FT (soft) 2023.05	-	-	-	-	-	95.5	-