Aspect Sentiment Triplet Extraction on Rest 2014 (test)
75.59F1 ScoreMiniConGTS
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MiniConGTSParadigm=Table tagging2026.05 | 75.59 | 76.1 | 75.08 | |
| D2E2SParadigm=Span-based2026.05 | 75.13 | 75.92 | 74.36 | |
| MvP + FiVeDVerification Model=DeepSeek-Reasoner, Paradigm=Generative2026.05 | 74.94 | 74.95 | 75.01 | |
| MvP + FiVeD-unifiedVerification Model=Qwen, Paradigm=Generative2026.05 | 74.67 | 73.36 | 76.04 | |
| MvP + FiVeDVerification Model=Qwen, Paradigm=Generative2026.05 | 74.6 | 73.65 | 75.59 | |
| MvP + FiVeD-unifiedVerification Model=DeepSeek-Reasoner, Paradigm=Generative2026.05 | 74.46 | 74.15 | 74.79 | |
| TAGSParadigm=Sequence tagging2026.05 | 74.36 | 74.92 | 73.81 | |
| SimSTARParadigm=Table tagging2026.05 | 73.86 | 76.23 | 71.63 | |
| LEGOParadigm=Generative2026.05 | 73.7 | — | — | |
| MvPParadigm=Generative2026.05 | 73.68 | 74.25 | 73.12 | |
| Paraphrase + FiVeD-unifiedVerification Model=DeepSeek-Reasoner, Paradigm=Generative2026.05 | 72.82 | 72.83 | 72.82 | |
| Paraphrase + FiVeD-unifiedVerification Model=Qwen, Paradigm=Generative2026.05 | 72.74 | 72.55 | 72.96 | |
| Paraphrase + FiVeDVerification Model=DeepSeek-Reasoner, Paradigm=Generative2026.05 | 72.48 | 72.2 | 72.8 | |
| Triple-MRCParadigm=MRC-based2026.05 | 72.45 | — | — | |
| Paraphrase + FiVeDVerification Model=Qwen, Paradigm=Generative2026.05 | 72.45 | 72.87 | 72.07 | |
| COM-MRCParadigm=MRC-based2026.05 | 72.01 | 75.46 | 68.91 | |
| GAS + FiVeDVerification Model=DeepSeek-Reasoner, Paradigm=Generative2026.05 | 71.91 | 68.36 | 75.86 | |
| Span-ASTEParadigm=Span-based2026.05 | 71.85 | 72.89 | 70.89 | |
| EMC-GCNParadigm=Table tagging2026.05 | 71.78 | 71.21 | 72.39 | |
| GAS + FiVeDVerification Model=Qwen, Paradigm=Generative2026.05 | 71.43 | 66.99 | 76.52 | |
| GAS + FiVeD-unifiedVerification Model=DeepSeek-Reasoner, Paradigm=Generative2026.05 | 71.4 | 66.41 | 77.23 | |
| GAS + FiVeD-unifiedVerification Model=Qwen, Paradigm=Generative2026.05 | 71.38 | 66.78 | 76.68 | |
| Dual-MRCParadigm=MRC-based2026.05 | 70.32 | 71.55 | 69.14 | |
| ParaphraseParadigm=Generative2026.05 | 70.12 | 71.03 | 66.98 | |
| GASParadigm=Generative2026.05 | 68.69 | 65.13 | 72.94 | |
| Span-BARTParadigm=Sequence tagging2026.05 | 65.25 | 65.52 | 64.99 | |
| GPT-4oEvaluation Prompting Strategy=few-shot, Paradigm=Generative2026.05 | 59.55 | 54.11 | 66.2 | |
| GPT-3.5-turboEvaluation Prompting Strategy=few-shot, Paradigm=Generative2026.05 | 57.55 | 51.51 | 65.19 | |
| GPT-3.5-turboEvaluation Prompting Strategy=CoT + few-shot, Paradigm=Generative2026.05 | 53.85 | 49.41 | 59.15 | |
| GPT-3.5-turboEvaluation Prompting Strategy=CoT, Paradigm=Generative2026.05 | 53.24 | 48.47 | 59.05 | |
| GPT-4oEvaluation Prompting Strategy=CoT + few-shot, Paradigm=Generative2026.05 | 52.54 | 46.81 | 59.86 | |
| Peng-two-stageParadigm=Sequence tagging2026.05 | 51.46 | 43.24 | 63.66 | |
| GPT-3.5-turboEvaluation Prompting Strategy=zero-shot, Paradigm=Generative2026.05 | 49.48 | 44.88 | 55.13 | |
| GPT-4oEvaluation Prompting Strategy=CoT, Paradigm=Generative2026.05 | 46.49 | 41.21 | 53.32 | |
| GPT-4oEvaluation Prompting Strategy=zero-shot, Paradigm=Generative2026.05 | 35.37 | 32.99 | 38.13 |