Aspect-Based Sentiment Analysis on Rest 14 (Accuracy, Macro-F1)
89.76AccuracyDABS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DABSSeeds=3, Evaluation protocol=Fine-tuning2026.05 | 89.76 | 84.87 | |
| DeBERTa+RCLEvaluation protocol=Fine-tuning2026.05 | 89.38 | 84.68 | |
| PConvRoBERTaEvaluation protocol=Fine-tuning2026.05 | 89.29 | 84.27 | |
| Flan-T5-base+Syn-ChainEvaluation protocol=Fine-tuning2026.05 | 88.39 | 82.79 | |
| ITGCNEvaluation protocol=Fine-tuning2026.05 | 88.21 | 82.35 | |
| LWEDA-ACT (2025)2025.09 | 88.11 | 82.12 | |
| RDGCN (2024)2025.09 | 87.49 | 81.16 | |
| CSADGCN (2024)2025.09 | 87.4 | 81.56 | |
| DSSK-GAN-BERTEvaluation protocol=Fine-tuning2026.05 | 87.32 | 82.35 | |
| SSEGCN (2022)2025.09 | 87.31 | 81.09 | |
| DC-GCNEvaluation protocol=Fine-tuning2026.05 | 87.29 | 81.48 | |
| IDGNN (2024)2025.09 | 87.25 | 81.16 | |
| OTESGN2025.09 | 87.21 | 80.47 | |
| KGAN-BERTEvaluation protocol=Fine-tuning2026.05 | 87.15 | 82.05 | |
| CADA (2025)2025.09 | 87.14 | 81.26 | |
| YORO (2024)2025.09 | 87.14 | 81.83 | |
| DualGCN (2021)2025.09 | 87.13 | 81.16 | |
| KDGN (2023)2025.09 | 87.01 | 81.94 | |
| APSCL (2023)2025.09 | 86.86 | 81.28 | |
| HPEP-GCN (2025)2025.09 | 86.86 | 80.65 | |
| SVCCL (2025)2025.09 | 86.34 | 80.16 | |
| DGEDT (2020)2025.09 | 86.3 | 80 | |
| BERT (2019)2025.09 | 85.97 | 80.09 | |
| CABiLSTM-BERTEvaluation protocol=Fine-tuning2026.05 | 83.75 | 75.87 | |
| ASHGATEvaluation protocol=Fine-tuning2026.05 | 83.45 | 76.55 | |
| Llama-3-8B-InstNumber of shots=5-shot2026.05 | 82.87 | 71.82 | |
| Qwen3-8BNumber of shots=5-shot2026.05 | 82.75 | 73.52 | |
| ASGCN (2019)2025.09 | 80.77 | 72.02 | |
| TNet (2018)2025.09 | 80.69 | 71.27 | |
| RAM (2017)2025.09 | 80.23 | 70.8 | |
| GPT-3.5 TurboNumber of shots=5-shot2026.05 | 78.85 | 66.5 |