GEC-DI (LM+GED)

Reported on 5 benchmarks across 1 task · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing5 results

Grammatical Error CorrectiononMuCGEC
F0.5· 2023-10-23
48.61
SOTA
Improving Seq2Seq Grammatical Error Correction via Decoding Interventions arXiv:2310.14534
Grammatical Error CorrectiononCoNLL-2014 Shared Task
F0.5· 2023-10-23
69.6
best: 72.8 (Ensembles of best 7 models + GRECO + GTP-rerank)
Improving Seq2Seq Grammatical Error Correction via Decoding Interventions arXiv:2310.14534
Grammatical Error CorrectiononCoNLL-2014 Shared Task
Precision· 2023-10-23
79.2
best: 83.9 (Ensembles of best 7 models + GRECO + GTP-rerank)
Improving Seq2Seq Grammatical Error Correction via Decoding Interventions arXiv:2310.14534
Grammatical Error CorrectiononCoNLL-2014 Shared Task
Recall· 2023-10-23
46.8
best: 53.8 (Unsupervised GEC + cLang8)
Improving Seq2Seq Grammatical Error Correction via Decoding Interventions arXiv:2310.14534
Grammatical Error CorrectiononBEA-2019 (test)
F0.5· 2023-10-23
73.1
best: 81.4 (Majority-voting ensemble on best 7 models)
Improving Seq2Seq Grammatical Error Correction via Decoding Interventions arXiv:2310.14534