Aspect-based Sentiment Analysis on SemEval Laptop 2014
85.16F1 ScoreTHOR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| THORModel Parameters=11B, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning2023.05 | 85.16 | — | — | — | |
| THORModel Parameters=11B, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning, Method Variation=w/o Reason-revising2023.05 | 84.83 | — | — | — | |
| THORModel Parameters=11B, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning, Method Variation=w/o Self-consistency2023.05 | 84.39 | — | — | — | |
| Flan-T5+PromptModel Parameters=11B, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning2023.05 | 82.44 | — | — | — | |
| THORModel Parameters=250M, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning2023.05 | 79.75 | — | — | — | |
| BERTAsp+SCAPTModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 79.15 | — | — | — | |
| Flan-T5+PromptModel Parameters=250M, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning2023.05 | 79.02 | — | — | — | |
| BERT+PromptModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 78.58 | — | — | — | |
| BERTAsp+CEPTModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 78.38 | — | — | — | |
| BERT+ISAIVModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 77.25 | — | — | — | |
| BERT+ADAModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 74.18 | — | — | — | |
| BERT+RGATModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 74.07 | — | — | — | |
| BERT+SPCModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 73.45 | — | — | — | |
| Extra2023.07 | 50.88 | 59.18 | 44.62 | — | |
| CoTAMshot_number (K)=10, generation_count (N)=3, backbone=RoBERTa-Large, embedding_model=SimCSE2023.07 | 38.26 | 33.33 | 44.9 | — | |
| FlipDA++2023.07 | 32.81 | 26.58 | 42.86 | — | |
| CoTDA2023.07 | 30.51 | 26.09 | 36.74 | — | |
| Base2023.07 | 25.93 | 23.73 | 28.57 | — | |
| LLM SL2023.07 | 14.09 | 18.56 | 22.73 | — | |
| ABSA-DEBERTa2024.05 | — | — | — | 82.76 | |
| Arctic-ABSA DecoderReasoning Capability=disabled2026.01 | — | — | — | 86.05 | |
| Arctic-ABSA Decoder-thinkingReasoning Capability=enabled2026.01 | — | — | — | 85.1 | |
| Arctic-ABSA EncoderReasoning Capability=disabled2026.01 | — | — | — | 83.99 | |
| Arctic-ABSA Encoder-thinkingReasoning Capability=enabled2026.01 | — | — | — | 85.1 | |
| Claude 3.5 SonnetModel Category=Closed-source2026.01 | — | — | — | 83.36 | |
| Dual-MRC2024.05 | — | — | — | 75.97 | |
| GPT-4oModel Category=Closed-source2026.01 | — | — | — | 83.36 | |
| InstructABSA12024.05 | — | — | — | 80.62 | |
| InstructABSA22024.05 | — | — | — | 81.56 | |
| IT-RER-ABSAk=42024.05 | — | — | — | 91.47 | |
| Llama3.1-405bModel Category=Open-source2026.01 | — | — | — | 82.41 | |
| Llama3.1-70bModel Category=Open-source2026.01 | — | — | — | 78.92 | |
| Llama3.1-8bModel Category=Open-source2026.01 | — | — | — | 73.22 | |
| Llama3.2-3bModel Category=Open-source2026.01 | — | — | — | 73.38 | |
| LSAT2024.05 | — | — | — | 86.31 | |
| Mistral Large 2Model Category=Open-source2026.01 | — | — | — | 80.98 | |
| RACL-BERT2024.05 | — | — | — | 73.91 | |
| SPAN2024.05 | — | — | — | 81.39 |