List your company
ResearchFormerly WizWand

Opinion Polarity Detection on MPQA

93.3AccuracyAdaSent
83.347285.931188.51591.0989Aug 25, 2014Jun 18, 2016Apr 13, 2018Feb 6, 2020Nov 30, 2021Sep 25, 2023Jul 20, 2025
Updated 6mo ago

Evaluation results

MethodLinks
AdaSent2015.0493.3
ADASENT2017.0493.3
FADS-ICLNumber of training shots (m)=4, LLM scale=1.5B2024.0591.1
BRNN2015.0490.3
LM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=LSTM2019.0590.15
RNN2015.0490.1
STM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=CNN2019.0589.95
NLM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=LSTM2019.0589.92
STM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=LSTM2019.0589.83
LSTM-400Emb.Size=4002019.0589.73
NLM+TSEDEmb.Size=50, Teacher Model Architecture=LSTM2019.0589.63
NLM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=CNN2019.0589.61
CNN-staticword vectors=static pre-trained (word2vec)2014.0889.6
CNN2015.0489.6
CNN2017.0489.6
STM+TSEDEmb.Size=50, Teacher Model Architecture=LSTM2019.0589.6
LM+TSEDEmb.Size=50, Teacher Model Architecture=LSTM2019.0589.59
CNN-non-staticword vectors=fine-tuned pre-trained (word2vec)2014.0889.5
LM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=CNN2019.0589.44
CNN-multichannelword vectors=pre-trained (word2vec), channels=22014.0889.4
SKIPTHOUGHTLayer Normalization=True2017.0489.3
LM+TSEDEmb.Size=50, Teacher Model Architecture=CNN2019.0589.14
STM+TSEDEmb.Size=50, Teacher Model Architecture=CNN2019.0589.13
ICLBackbone=Llama-2 70B, Shots=128-shots2024.0589.1
NLM+TSEDEmb.Size=50, Teacher Model Architecture=CNN2019.0588.9
CNN-400Emb.Size=4002019.0588.78
IS-BERT_baseBackbone=BERT-base2022.0388.75
SCD-BERT_baseBackbone=BERT-base2022.0388.67
Avg. BERT embeddingsBackbone=BERT-base, Pooling=Avg2022.0388.66
ENCEmb.Size=502019.0588.65
SimCSE-BERT_baseBackbone=BERT-base2022.0388.6
BYTE MLSTMClassifier=Logistic Regression, Penalty=L12017.0488.5
BERT [CLS]-embeddingBackbone=BERT-base, Pooling=CLS2022.0388.21
SCD-RoBERTa_baseBackbone=RoBERTa-base2022.0388.19
USE_T+DAN (w2v w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=DAN2018.0388.14
FADS-ICLNumber of training shots (m)=32, LLM scale=1.5B2024.0588
FADS-ICLBackbone=Llama-2 13B, Shots=128-shots2024.0588
SKIPTHOUGHT2017.0487.9
GloVe embeddings (avg.)Model=GloVe (avg.)2022.0387.85
FADS-ICLBackbone=Llama-2 70B, Shots=128-shots2024.0587.7
GPT Large FTBackbone=GPT Large, Shots=N/A (Fine-tuned)2024.0587.6
ICLBackbone=Llama-1 30B, Shots=128-shots2024.0587.6
USE_T+CNN (w2v w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=CNN2018.0387.32
FADS-ICLBackbone=Llama-2 7B, Shots=128-shots2024.0587.3
FADS-ICLBackbone=Llama-1 30B, Shots=128-shots2024.0587.3
CCAE2014.0887.2
DictRep BOW+embs.Data Source Type=Other structured data resource2016.0287.2
USE_T+DAN (lrn w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=DAN2018.0387.14
SkipThoughtData Source Type=Ordered Sentences2016.0287.1
Skip-thoughtModel=Skip-thought2022.0387.1
USE_D+DAN (w2v w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=DAN2018.0387.01
ICLBackbone=Llama-2 7B, Shots=128-shots2024.0587
USE TUniversal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=None, Transfer Task Model=None2018.0386.98
SDAE+embs.Data Source Type=Unordered Sentences2016.0286.9
SDAE2017.0486.9
USE_T+CNN (lrn w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=CNN2018.0386.85
ICLBackbone=Llama-1 7B, Shots=128-shots2024.0586.8
SimCSE-RoBERTa_baseBackbone=RoBERTa-base2022.0386.63
RAE2014.0886.4
RAE2015.0486.4
CBOW2015.0486.4
CNN-50Emb.Size=502019.0586.4
Sent-Parser2014.0886.3
NBSVM2014.0886.3
MNB2014.0886.3
F-Dropout2014.0886.3
NB-SVM2015.0486.3
MNB2015.0486.3
NBSVM2017.0486.3
SAE+embs.Data Source Type=Unordered Sentences2016.0286.2
CPHRASEData Source Type=2.8B words2016.0286.2
G-Dropout2014.0886.1
Tree-CRF2014.0886.1
USE_D+DAN (lrn w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=DAN2018.0385.97
FADS-ICLBackbone=Llama-1 13B, Shots=128-shots2024.0585.9
USE_D+CNN (w2v w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=CNN2018.0385.87
FADS-ICLBackbone=Llama-1 7B, Shots=128-shots2024.0585.8
LSTM-50Emb.Size=502019.0585.66
USE_D+CNN (lrn w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=CNN2018.0385.53
ICLBackbone=Llama-1 13B, Shots=128-shots2024.0585.5
ICLBackbone=Llama-2 13B, Shots=128-shots2024.0585.5
kNN-promptBackbone=Llama-2 13B, Shots=128-shots2024.0585.5
USE DUniversal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=None, Transfer Task Model=None2018.0385.38
FADS-ICLNumber of training shots (m)=128, LLM scale=1.5B2024.0584.8
FADS-ICLBackbone=GPT-2 1.5B, Shots=128-shots2024.0584.8
FADS-ICLcandidate pool size (m)=1282024.0584.8
kNN-promptingBackbone=Llama-1 7B, Shots=128-shots2024.0584.6
GrConv2015.0484.5
kNN-promptingBackbone=Llama-2 13B, Shots=128-shots2024.0584.5
kNN-promptingBackbone=Llama-1 13B, Shots=128-shots2024.0584.4
kNN-promptBackbone=Llama-1 13B, Shots=128-shots2024.0584.3
kNN-promptBackbone=Llama-1 30B, Shots=128-shots2024.0584.3
kNN-promptingBackbone=Llama-1 30B, Shots=128-shots2024.0584.3
kNN-promptBackbone=Llama-2 70B, Shots=128-shots2024.0584.3
FADS-ICLcandidate pool size (m)=2562024.0584.2
RoBERTa [CLS]-embeddingBackbone=RoBERTa-base, Pooling=CLS2022.0384.18
kNN-promptingBackbone=Llama-2 70B, Shots=128-shots2024.0584.1
ICLNumber of training shots (m)=16, LLM scale=1.5B2024.0583.9
kNN-promptingNumber of training shots (m)=32, LLM scale=1.5B2024.0583.9
GNN2025.0783.73