# Natural Language Understanding on MRPC, QQP, STS-B, MNLI, QNLI, RTE, SNLI, Sci, SICK, and Twi Source: [Natural Language Understanding on MRPC, QQP, STS-B, MNLI, QNLI, RTE, SNLI, Sci, SICK, and Twi]() Collection index: [llms.txt]() Data updated: 2026-06-05T07:36:41.804Z Aggregate evaluation of model performance across eleven specific Natural Language Understanding tasks: MRPC, QQP, STS-B, MNLI (m/mm), QNLI, RTE, SNLI, Sci, SICK, and Twi. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Natural Language Understanding | | Dataset (including any stated split) | Ten Benchmark Datasets MRPC, QQP, STS-B, MNLI, QNLI, RTE, SNLI, Sci, SICK, Twi | | Metric | MRPC Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [MRPC Accuracy](): Higher is better; unit: PERCENT. - [QQP Accuracy](): Higher is better; unit: PERCENT. - [STS-B Score](): Higher is better; unit: PERCENT. - [MNLI (matched) Accuracy](): Higher is better; unit: PERCENT. - [MNLI (mismatched) Accuracy](): Higher is better; unit: PERCENT. - [QNLI Accuracy](): Higher is better; unit: PERCENT. - [RTE Accuracy](): Higher is better; unit: PERCENT. - [SNLI Accuracy](): Higher is better; unit: PERCENT. - [Sci Accuracy](): Higher is better; unit: PERCENT. - [SICK Accuracy](): Higher is better; unit: PERCENT. - [Twi Accuracy](): Higher is better; unit: PERCENT. - [Average NLU Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | MRPC Accuracy | Result date | Report paper | | --- | --- | --- | --- | | RoBERTa-Large-MGRN | 90.2 | Not recorded | 6a22584686173ad9c146882a | | BERT-Large-MGRN | 89.9 | Not recorded | 6a22584686173ad9c146882a | | RoBERTa-Base-MGRN | 89.5 | Not recorded | 6a22584686173ad9c146882a | | RoBERTa-Large | 89.4 | Not recorded | 6a22584686173ad9c146882a | | BERT-Base-MGRN | 89.3 | Not recorded | 6a22584686173ad9c146882a | | RoBERTa-Base | 89.3 | Not recorded | 6a22584686173ad9c146882a | | SyntaxBERT | 89.2 | Not recorded | 6a22584686173ad9c146882a | | MGRN | 89.1 | Not recorded | 6a22584686173ad9c146882a | | BERT-Large | 88.9 | Not recorded | 6a22584686173ad9c146882a | | UERBERT | 88.3 | Not recorded | 6a22584686173ad9c146882a | | SemBERT | 88.2 | Not recorded | 6a22584686173ad9c146882a | | BERT-Base | 87.2 | Not recorded | 6a22584686173ad9c146882a | | BiLSTM+ELMo+Attnt | 84.6 | Not recorded | 6a22584686173ad9c146882a | | CAFE | 82.4 | Not recorded | 6a22584686173ad9c146882a | | OpenAI GPT | 82.3 | Not recorded | 6a22584686173ad9c146882a | | Transformer | 81.7 | Not recorded | 6a22584686173ad9c146882a | | ESIM | 80.3 | Not recorded | 6a22584686173ad9c146882a | | BiMPM | 79.6 | Not recorded | 6a22584686173ad9c146882a | ### RoBERTa-Large-MGRN MRPC Accuracy: 90.2. Result ID: 6a227c6ffbc373592c20214e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### BERT-Large-MGRN MRPC Accuracy: 89.9. Result ID: 6a227c6ffbc373592c20214a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### RoBERTa-Base-MGRN MRPC Accuracy: 89.5. Result ID: 6a227c6ffbc373592c20214c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### RoBERTa-Large MRPC Accuracy: 89.4. Result ID: 6a227c6ffbc373592c20214d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### BERT-Base-MGRN MRPC Accuracy: 89.3. Result ID: 6a227c6ffbc373592c202148. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### RoBERTa-Base MRPC Accuracy: 89.3. Result ID: 6a227c6ffbc373592c20214b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### SyntaxBERT MRPC Accuracy: 89.2. Result ID: 6a227c6ffbc373592c202145. [Reporting paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### MGRN MRPC Accuracy: 89.1. Result ID: 6a227c6ffbc373592c202146. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### BERT-Large MRPC Accuracy: 88.9. Result ID: 6a227c6ffbc373592c202149. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### UERBERT MRPC Accuracy: 88.3. Result ID: 6a227c6ffbc373592c202143. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### SemBERT MRPC Accuracy: 88.2. Result ID: 6a227c6ffbc373592c202144. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### BERT-Base MRPC Accuracy: 87.2. Result ID: 6a227c6ffbc373592c202147. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### BiLSTM+ELMo+Attnt MRPC Accuracy: 84.6. Result ID: 6a227c6ffbc373592c202141. [Reporting paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### CAFE MRPC Accuracy: 82.4. Result ID: 6a227c6ffbc373592c20213e. [Reporting paper]() Recorded evaluation configuration: {"Pre-train":"false"}. This record does not assert independent reproduction. ### OpenAI GPT MRPC Accuracy: 82.3. Result ID: 6a227c6ffbc373592c202142. [Reporting paper]() Recorded evaluation configuration: {"Pre-train":"true"}. This record does not assert independent reproduction. ### Transformer MRPC Accuracy: 81.7. Result ID: 6a227c6ffbc373592c202140. [Reporting paper]() Recorded evaluation configuration: {"Pre-train":"false"}. This record does not assert independent reproduction. ### ESIM MRPC Accuracy: 80.3. Result ID: 6a227c6ffbc373592c20213f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-train":"false"}. This record does not assert independent reproduction. ### BiMPM MRPC Accuracy: 79.6. Result ID: 6a227c6ffbc373592c20213d. [Reporting paper]() Recorded evaluation configuration: {"Pre-train":"false"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.