Aspect Sentiment Quad Prediction on Rest16
66.49F1 ScoreLLM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LLMMode=Fine-tuned2026.05 | 66.49 | — | — | |
| Gemma 4Size=31B, Mode=Fine-tuned2026.05 | 66.49 | — | — | |
| UGTS2026.05 | 65.1 | — | — | |
| E4LTraining protocol=State-of-the-art2025.11 | 64.21 | 63.98 | 64.46 | |
| E4Lcollaborative=true2025.11 | 64.21 | — | — | |
| MUL+Scorer&ReRank (AI)collaborative=true, refinement=Scorer & ReRank (AI)2025.11 | 63.88 | — | — | |
| GAS+Scorer&ReRank (AI)collaborative=true, refinement=Scorer & ReRank (AI)2025.11 | 63.51 | — | — | |
| BvSPLearning mode=Full-shot, top-k=152024.06 | 63.49 | 68.16 | 59.42 | |
| SimRPTraining protocol=State-of-the-art2025.11 | 63.42 | 62.74 | 64.12 | |
| G2CBackbone=T5-base2026.03 | 62.8 | 61.24 | 64.48 | |
| E2TP2026.05 | 62.57 | — | — | |
| LLM-MvPShots=1002026.05 | 62.51 | — | — | |
| LLM-MvP# Shots=1002026.05 | 62.51 | — | — | |
| LLM-MvPSetting=100-shot2026.05 | 62.51 | — | — | |
| SCRAPcollaborative=true2025.11 | 62.48 | — | — | |
| SCRAP (2024)2026.03 | 62.48 | 69.59 | 56.7 | |
| G2C (w/o identify pairs)Backbone=T5-base, Variant=w/o identify pairs2026.03 | 61.83 | 60.29 | 63.45 | |
| STAR (2024)2026.03 | 61.7 | 60.54 | 62.9 | |
| GDP (2024)2026.03 | 61.61 | 61.16 | 62.08 | |
| BvSPLearning mode=Full-shot, top-k=32024.06 | 61.4 | 63.59 | 59.35 | |
| G2C (w/o Corrector Stage-1)Backbone=T5-base, Variant=w/o Corrector2026.03 | 61.33 | 59.9 | 62.83 | |
| DOT (2025)2026.03 | 61.24 | — | — | |
| IVLSTraining protocol=State-of-the-art2025.11 | 61.04 | 62.69 | 59.75 | |
| GenDATraining protocol=State-of-the-art2025.11 | 60.88 | 60.08 | 61.7 | |
| Orca 2Size=13B, Mode=Fine-tuned2026.05 | 60.82 | — | — | |
| DLO+UAUL (2023)2026.03 | 60.5 | 59.02 | 62.05 | |
| MULTraining protocol=State-of-the-art2025.11 | 60.47 | 59.24 | 61.75 | |
| Special_Symbols+UAUL (2023)2026.03 | 60.47 | 59.24 | 61.75 | |
| MvPLearning mode=Full-shot, top-k=152024.06 | 60.39 | — | — | |
| MvPTraining protocol=State-of-the-art2025.11 | 60.39 | — | — | |
| MVP (2023)2026.03 | 60.39 | — | — | |
| MvPMode=Fine-tuned2026.05 | 60.39 | — | — | |
| MvP2026.05 | 60.39 | — | — | |
| OTCLTraining protocol=State-of-the-art2025.11 | 60.11 | 58.31 | 62.02 | |
| DLOLearning mode=Full-shot2024.06 | 59.79 | 57.92 | 61.8 | |
| DLOTraining protocol=State-of-the-art2025.11 | 59.79 | 57.92 | 61.8 | |
| DLO (2022)2026.03 | 59.79 | 57.92 | 61.8 | |
| LLM-MvP# Shots=502026.05 | 59.65 | — | — | |
| MvPLearning mode=Full-shot, top-k=32024.06 | 59.62 | 60.27 | 58.94 | |
| Special_Symbols (2022)2026.03 | 59.53 | 58.74 | 60.35 | |
| ILOLearning mode=Full-shot2024.06 | 59.32 | 57.58 | 61.17 | |
| ILOTraining protocol=State-of-the-art2025.11 | 59.32 | 57.58 | 61.17 | |
| ILO (2022)2026.03 | 59.32 | 57.58 | 61.17 | |
| Self-consistency# Shots=1002026.05 | 58.79 | — | — | |
| Orca 2Size=7B, Mode=Fine-tuned2026.05 | 58.63 | — | — | |
| ParaphraseLearning mode=Full-shot2024.06 | 57.93 | 56.63 | 59.3 | |
| ParaphraseTraining protocol=State-of-the-art2025.11 | 57.93 | 56.63 | 59.3 | |
| Paraphrase (2021)2026.03 | 57.93 | 56.63 | 59.3 | |
| Para.Mode=Fine-tuned2026.05 | 57.93 | — | — | |
| Paraphrase2026.05 | 57.93 | — | — | |
| LEGO-ABSATraining protocol=State-of-the-art2025.11 | 57.7 | — | — | |
| GASTraining protocol=State-of-the-art2025.11 | 57.55 | 57.3 | 57.82 | |
| Self-consistency# Shots=502026.05 | 56.37 | — | — | |
| Single-order# Shots=1002026.05 | 56.36 | — | — | |
| GASLearning mode=Full-shot2024.06 | 56.04 | 54.54 | 57.62 | |
| GAS (2021)2026.03 | 56.04 | 54.54 | 57.62 | |
| GAS2026.05 | 56.04 | — | — | |
| Single-order# Shots=502026.05 | 54.7 | — | — | |
| Gemini 2.0 FlashShots=1002026.05 | 54.4 | — | — | |
| LLM-MvP# Shots=Avg.2026.05 | 54.32 | — | — | |
| LLM-MvPShots=102026.05 | 54 | — | — | |
| LLM-MvP# Shots=102026.05 | 54 | — | — | |
| Self-consistency# Shots=102026.05 | 51.7 | — | — | |
| Self-consistency# Shots=Avg.2026.05 | 50.35 | — | — | |
| Single-order# Shots=102026.05 | 48.98 | — | — | |
| Single-order# Shots=Avg.2026.05 | 47.54 | — | — | |
| gpt-oss-120BShots=100, Prompting Strategy=CoT2026.05 | 46.62 | — | — | |
| Reasoning# Shots=1002026.05 | 46.62 | — | — | |
| Gemma-3-27BShots=10, Prompting Strategy=SC2026.05 | 46.23 | — | — | |
| GPT-4oTraining protocol=Training-free, Shot count=10-shot2025.11 | 45 | — | — | |
| Reasoning# Shots=502026.05 | 44.67 | — | — | |
| Extract-Classify (2021)2026.03 | 43.88 | 38.4 | 50.93 | |
| Extract-ClassifyTraining protocol=State-of-the-art2025.11 | 43.77 | 38.4 | 50.93 | |
| Reasoning# Shots=Avg.2026.05 | 42.9 | — | — | |
| gpt-oss-120BShots=10, Prompting Strategy=CoT2026.05 | 42.05 | — | — | |
| Reasoning# Shots=102026.05 | 42.05 | — | — | |
| LLM-MvPShots=02026.05 | 41.15 | — | — | |
| LLM-MvP# Shots=02026.05 | 41.15 | — | — | |
| ChatGPTTraining protocol=Training-free, Shot count=few-shot2025.11 | 40.81 | 36.09 | 46.93 | |
| gpt-3.5-turboShots=102026.05 | 40.15 | — | — | |
| gpt-oss-120BShots=0, Prompting Strategy=CoT2026.05 | 38.27 | — | — | |
| Reasoning# Shots=02026.05 | 38.27 | — | — | |
| Llama-3.1-70BTraining protocol=Training-free, Shot count=10-shot2025.11 | 37.1 | — | — | |
| Self-consistency# Shots=02026.05 | 34.54 | — | — | |
| gpt4o-miniShots=02026.05 | 34.31 | — | — | |
| ChatABSAShots=102026.05 | 33.26 | — | — | |
| ChatABSAShots=02026.05 | 30.42 | — | — | |
| Single-order# Shots=02026.05 | 30.13 | — | — | |
| Gemma-3-27BShots=0, Prompting Strategy=SC2026.05 | 28.96 | — | — | |
| HGCN-BERT+BERT-TFM(2021)2026.03 | 26.9 | 27.4 | 26.41 | |
| gpt4o-miniShots=0, Prompting Strategy=CoT2026.05 | 26.73 | — | — | |
| HGCN-BERT+BERT-Lineear(2021)2026.03 | 24.68 | 25.36 | 24.03 | |
| gpt-3.5-turboShots=02026.05 | 14.02 | — | — |