Semantic Role Labeling on CoNLL WSJ English benchmark 2009 (test)
92.83F1 ScoreFei et al. (2021a) ++ROBERTa
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Fei et al. (2021a) ++ROBERTaPredicates=With pre-identified, Backbone=RoBERTa2025.06 | 92.83 | 92.89 | 92.8 | |
| Fei et al. (2021a)Predicates=With pre-identified2025.06 | 92.45 | 92.24 | 92.53 | |
| Llama3-8B (Fine-tune)Predicates=With pre-identified, Model Type=Fine-tuned LLM2025.06 | 91.89 | 92.78 | 91.02 | |
| Li et al. (2020) + BERTPredicates=With pre-identified, Backbone=BERT2025.06 | 91.77 | 92.59 | 90.98 | |
| Fernández-González (2023) + BERTPredicates=With pre-identified, Backbone=BERT2025.06 | 91.4 | 91.3 | 91.6 | |
| UniSRLModel Type=Single, Predicate Setting=Pre-identified2019.01 | 90.4 | 89.6 | 91.2 | |
| Fernández-González (2023)Predicates=With pre-identified2025.06 | 90.4 | 90.2 | 90.5 | |
| Li et al. (2020)Predicates=With pre-identified2025.06 | 90.26 | 91.6 | 88.95 | |
| Li et al. (2018b)Model Type=Single, Predicate Setting=Pre-identified2019.01 | 89.8 | 90.3 | 89.3 | |
| Cai et al. (2018)Model Type=Single, Predicate Setting=Pre-identified2019.01 | 89.6 | 89.9 | 89.2 | |
| He et al. (2018b)Model Type=Single, Predicate Setting=Pre-identified2019.01 | 89.5 | 89.7 | 89.3 | |
| Marcheggiani and Titov (2017)Model Type=Ensemble, Predicate Setting=Pre-identified2019.01 | 89.1 | 90.5 | 87.7 | |
| Llama3-8B (Fine-tune)Predicates=Without pre-identified, Model Type=Fine-tuned LLM2025.06 | 89.07 | 89.92 | 88.23 | |
| Zhou et al. (2020) + BERTPredicates=With pre-identified, Backbone=BERT2025.06 | 88.91 | 89.04 | 88.79 | |
| Li et al. (2020) + BERTPredicates=Without pre-identified, Backbone=BERT2025.06 | 88.7 | 88.77 | 88.62 | |
| Fernández-González (2023) + BERTPredicates=Without pre-identified, Backbone=BERT2025.06 | 88.5 | 87.2 | 89.8 | |
| Marcheggiani and Titov (2017)Model Type=Single, Predicate Setting=Pre-identified2019.01 | 88 | 89.1 | 86.8 | |
| Roth and Lapata (2016)Model Type=Ensemble, Predicate Setting=Pre-identified2019.01 | 87.9 | 90.3 | 85.7 | |
| Roth and Lapata (2016) (Global)Model Type=Single, Predicate Setting=Pre-identified2019.01 | 87.7 | 90 | 85.5 | |
| Marcheggiani et al. (2017)Model Type=Single, Predicate Setting=Pre-identified2019.01 | 87.7 | 88.7 | 86.8 | |
| FitzGerald et al. (2015)Model Type=Ensemble, Predicate Setting=Pre-identified2019.01 | 87.7 | — | — | |
| Zhou et al. (2020) + BERTPredicates=Without pre-identified, Backbone=BERT2025.06 | 87.62 | 86.77 | 88.49 | |
| FitzGerald et al. (2015) (Struct.)Model Type=Single, Predicate Setting=Pre-identified2019.01 | 87.3 | — | — | |
| Fernández-González (2023)Predicates=Without pre-identified2025.06 | 86.9 | 85.9 | 88 | |
| Zhao et al. (2009)Model Type=Single, Predicate Setting=Pre-identified2019.01 | 86.2 | — | — | |
| Li et al. (2020)Predicates=Without pre-identified2025.06 | 85.86 | 86.16 | 85.56 | |
| Zhou et al. (2020)Predicates=Without pre-identified2025.06 | 85.86 | 84.24 | 87.55 | |
| Zhou et al. (2020)Predicates=With pre-identified2025.06 | 85.84 | 85.93 | 85.76 | |
| ChatGPT+SimCSE KNNPredicates=With pre-identified, Reference=Sun et al. (2023)2025.06 | 84.8 | — | — | |
| Llama3-8B (Frozen)Predicates=With pre-identified, Model Type=Frozen LLM2025.06 | 4.43 | 4.19 | 4.71 | |
| Llama3-8B (Frozen)Predicates=Without pre-identified, Model Type=Frozen LLM2025.06 | 2.17 | 2.63 | 1.85 |