| chinese-named-entity-recognition-on-msra | ERNIE 2.0 Large | #10 | F1: 95 |
| chinese-named-entity-recognition-on-msra | ERNIE 2.0 Base | #15 | F1: 93.8 |
| linguistic-acceptability-on-cola | ERNIE 2.0 Large | #19 | Accuracy: 63.5% |
| linguistic-acceptability-on-cola | ERNIE 2.0 Base | #26 | Accuracy: 55.2% |
| natural-language-inference-on-multinli | ERNIE 2.0 Large | #11 | Matched: 88.7Mismatched: 88.8 |
| natural-language-inference-on-multinli | ERNIE 2.0 Base | #23 | Matched: 86.1Mismatched: 85.5 |
| natural-language-inference-on-qnli | ERNIE 2.0 Large | #12 | Accuracy: 94.6% |
| natural-language-inference-on-qnli | ERNIE 2.0 Base | #19 | Accuracy: 92.9% |
| natural-language-inference-on-rte | ERNIE 2.0 Large | #30 | Accuracy: 80.2% |
| natural-language-inference-on-rte | ERNIE 2.0 Base | #39 | Accuracy: 74.8% |
| natural-language-inference-on-wnli | ERNIE 2.0 Large | #17 | Accuracy: 67.8 |
| question-answering-on-quora-question-pairs | ERNIE 2.0 Large | #6 | Accuracy: 90.1% |
| question-answering-on-quora-question-pairs | ERNIE 2.0 Base | #9 | Accuracy: 89.8% |
| semantic-textual-similarity-on-mrpc | ERNIE 2.0 Large | #22 | Accuracy: 87.4% |
| semantic-textual-similarity-on-mrpc | ERNIE 2.0 Base | #28 | Accuracy: 86.1% |
| semantic-textual-similarity-on-sts-benchmark | ERNIE 2.0 Large | #9 | Pearson Correlation: 0.912 |
| semantic-textual-similarity-on-sts-benchmark | ERNIE 2.0 Base | #17 | Pearson Correlation: 0.876 |
| sentiment-analysis-on-sst-2-binary | ERNIE 2.0 Base | #24 | Accuracy: 95 |