RoBERTa (single model)

Reported on 3 benchmarks across 2 tasks · 2 papers

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

Natural Language Processing4 results

Question AnsweringonSQuAD2.0
EM· 2019-07-26
86.82
best: 90.939 (IE-Net (ensemble))
RoBERTa: A Robustly Optimized BERT Pretraining Approach arXiv:1907.11692
Question AnsweringonSQuAD2.0
F1· 2019-07-26
89.795
best: 93.214 (IE-Net (ensemble))
RoBERTa: A Robustly Optimized BERT Pretraining Approach arXiv:1907.11692
Question AnsweringonSQuAD2.0
EM· 2019-07-26
86.82
best: 90.939 (IE-Net (ensemble))
RoBERTa: A Robustly Optimized BERT Pretraining Approach arXiv:1907.11692
Question AnsweringonSQuAD2.0
F1· 2019-07-26
89.795
best: 93.214 (IE-Net (ensemble))
RoBERTa: A Robustly Optimized BERT Pretraining Approach arXiv:1907.11692

Adversarial1 result

Adversarial RobustnessonAdvGLUE
Accuracy· 2021-11-04
0.5021
best: 0.6086 (DeBERTa (single model))
Adversarial GLUE: A Multi-Task Benchmark for Robustness Evaluation of Language Models arXiv:2111.02840