Aspect Sentiment Quad Prediction on Rest15
56.78F1 ScoreLLM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LLMMode=Fine-tuned2026.05 | 56.78 | — | — | |
| Gemma 4Size=31B, Mode=Fine-tuned2026.05 | 56.78 | — | — | |
| LLM-MvPShots=1002026.05 | 54.94 | — | — | |
| LLM-MvP# Shots=1002026.05 | 54.94 | — | — | |
| LLM-MvPSetting=100-shot2026.05 | 54.94 | — | — | |
| E4LTraining protocol=State-of-the-art2025.11 | 54.73 | 54.12 | 55.35 | |
| E4Lcollaborative=true2025.11 | 54.73 | — | — | |
| SimRPTraining protocol=State-of-the-art2025.11 | 53.3 | 53.12 | 53.5 | |
| BvSPLearning mode=Full-shot, top-k=152024.06 | 53.17 | 60.96 | 47.15 | |
| UGTS2026.05 | 52.59 | — | — | |
| G2CBackbone=T5-base2026.03 | 52.4 | 51.26 | 53.38 | |
| Orca 2Size=13B, Mode=Fine-tuned2026.05 | 52.29 | — | — | |
| Self-consistency# Shots=1002026.05 | 52.07 | — | — | |
| LLM-MvP# Shots=502026.05 | 52.07 | — | — | |
| MUL+Scorer&ReRank (AI)collaborative=true, refinement=Scorer & ReRank (AI)2025.11 | 51.97 | — | — | |
| E2TP2026.05 | 51.94 | — | — | |
| DOT (2025)2026.03 | 51.91 | — | — | |
| GAS+Scorer&ReRank (AI)collaborative=true, refinement=Scorer & ReRank (AI)2025.11 | 51.74 | — | — | |
| Orca 2Size=7B, Mode=Fine-tuned2026.05 | 51.5 | — | — | |
| STAR (2024)2026.03 | 51.37 | 50.8 | 51.95 | |
| IVLSTraining protocol=State-of-the-art2025.11 | 51.28 | 54.46 | 48.53 | |
| MvPLearning mode=Full-shot, top-k=152024.06 | 51.04 | — | — | |
| MvPTraining protocol=State-of-the-art2025.11 | 51.04 | — | — | |
| MVP (2023)2026.03 | 51.04 | — | — | |
| MvPMode=Fine-tuned2026.05 | 51.04 | — | — | |
| MvP2026.05 | 51.04 | — | — | |
| BvSPLearning mode=Full-shot, top-k=32024.06 | 50.83 | 54.63 | 47.53 | |
| Gemini 2.0 FlashShots=1002026.05 | 50.7 | — | — | |
| Single-order# Shots=1002026.05 | 50.7 | — | — | |
| G2C (w/o identify pairs)Backbone=T5-base, Variant=w/o identify pairs2026.03 | 50.15 | 49.15 | 51.19 | |
| GenDATraining protocol=State-of-the-art2025.11 | 50.01 | 49.74 | 50.29 | |
| SCRAPcollaborative=true2025.11 | 49.93 | — | — | |
| SCRAP (2024)2026.03 | 49.93 | 55.45 | 45.41 | |
| MULTraining protocol=State-of-the-art2025.11 | 49.75 | 49.12 | 50.39 | |
| Special_Symbols+UAUL (2023)2026.03 | 49.75 | 49.12 | 50.39 | |
| GDP (2024)2026.03 | 49.75 | 49.2 | 50.31 | |
| G2C (w/o Corrector Stage-1)Backbone=T5-base, Variant=w/o Corrector2026.03 | 49.75 | 48.62 | 50.94 | |
| Self-consistency# Shots=502026.05 | 49.75 | — | — | |
| MvPLearning mode=Full-shot, top-k=32024.06 | 49.32 | 49.59 | 48.93 | |
| OTCLTraining protocol=State-of-the-art2025.11 | 49.27 | 47.86 | 50.77 | |
| DLO+UAUL (2023)2026.03 | 49.26 | 48.03 | 50.54 | |
| ILOLearning mode=Full-shot2024.06 | 49.05 | 47.78 | 50.38 | |
| ILOTraining protocol=State-of-the-art2025.11 | 49.05 | 47.78 | 50.38 | |
| ILO (2022)2026.03 | 49.05 | 47.78 | 50.38 | |
| Special_Symbols (2022)2026.03 | 48.58 | 48.24 | 48.93 | |
| DLOLearning mode=Full-shot2024.06 | 48.18 | 47.08 | 49.33 | |
| DLOTraining protocol=State-of-the-art2025.11 | 48.18 | 47.08 | 49.33 | |
| DLO (2022)2026.03 | 48.18 | 47.08 | 49.33 | |
| Single-order# Shots=502026.05 | 47.99 | — | — | |
| ParaphraseLearning mode=Full-shot2024.06 | 46.93 | 46.16 | 47.72 | |
| ParaphraseTraining protocol=State-of-the-art2025.11 | 46.93 | 46.16 | 47.72 | |
| Paraphrase (2021)2026.03 | 46.93 | 46.16 | 47.72 | |
| Para.Mode=Fine-tuned2026.05 | 46.93 | — | — | |
| Paraphrase2026.05 | 46.93 | — | — | |
| LLM-MvP# Shots=Avg.2026.05 | 46.89 | — | — | |
| GASTraining protocol=State-of-the-art2025.11 | 46.57 | 47.15 | 46.01 | |
| Extract-Classify (2021)2026.03 | 46.42 | 35.64 | 37.25 | |
| LLM-MvPShots=102026.05 | 46.28 | — | — | |
| LLM-MvP# Shots=102026.05 | 46.28 | — | — | |
| GASLearning mode=Full-shot2024.06 | 45.98 | 45.31 | 46.7 | |
| GAS (2021)2026.03 | 45.98 | 45.31 | 46.7 | |
| GAS2026.05 | 45.98 | — | — | |
| LEGO-ABSATraining protocol=State-of-the-art2025.11 | 45.8 | — | — | |
| Self-consistency# Shots=Avg.2026.05 | 43.04 | — | — | |
| Self-consistency# Shots=102026.05 | 42.13 | — | — | |
| Single-order# Shots=Avg.2026.05 | 41.18 | — | — | |
| Single-order# Shots=102026.05 | 40.43 | — | — | |
| Gemma-3-27BShots=10, Prompting Strategy=SC2026.05 | 39.95 | — | — | |
| gpt-oss-120BShots=100, Prompting Strategy=CoT2026.05 | 39.95 | — | — | |
| Reasoning# Shots=1002026.05 | 39.95 | — | — | |
| Reasoning# Shots=502026.05 | 37.51 | — | — | |
| GPT-4oTraining protocol=Training-free, Shot count=10-shot2025.11 | 37.08 | — | — | |
| Reasoning# Shots=Avg.2026.05 | 36.79 | — | — | |
| Extract-ClassifyTraining protocol=State-of-the-art2025.11 | 36.42 | 35.64 | 37.25 | |
| gpt-oss-120BShots=10, Prompting Strategy=CoT2026.05 | 36.39 | — | — | |
| Reasoning# Shots=102026.05 | 36.39 | — | — | |
| LLM-MvPShots=02026.05 | 34.28 | — | — | |
| LLM-MvP# Shots=02026.05 | 34.28 | — | — | |
| gpt-oss-120BShots=0, Prompting Strategy=CoT2026.05 | 33.33 | — | — | |
| Reasoning# Shots=02026.05 | 33.33 | — | — | |
| ChatGPTTraining protocol=Training-free, Shot count=few-shot2025.11 | 33.26 | 29.66 | 37.86 | |
| Llama-3.1-70BTraining protocol=Training-free, Shot count=10-shot2025.11 | 32.84 | — | — | |
| ChatABSAShots=102026.05 | 32.14 | — | — | |
| gpt-3.5-turboShots=102026.05 | 30.92 | — | — | |
| Self-consistency# Shots=02026.05 | 28.21 | — | — | |
| ChatABSAShots=02026.05 | 27.11 | — | — | |
| Single-order# Shots=02026.05 | 25.59 | — | — | |
| gpt4o-miniShots=02026.05 | 25.24 | — | — | |
| Gemma-3-27BShots=0, Prompting Strategy=SC2026.05 | 24.73 | — | — | |
| HGCN-BERT+BERT-TFM(2021)2026.03 | 23.65 | 25.55 | 22.01 | |
| HGCN-BERT+BERT-Lineear(2021)2026.03 | 22.15 | 24.32 | 20.25 | |
| gpt4o-miniShots=0, Prompting Strategy=CoT2026.05 | 21.55 | — | — | |
| gpt-3.5-turboShots=02026.05 | 10.46 | — | — |