# Natural Language Understanding on GLUE (test) (Subset: MNLI, SST-2, CoLA, QNLI, MRPC) Source: [Natural Language Understanding on GLUE (test) (Subset: MNLI, SST-2, CoLA, QNLI, MRPC)]() Collection index: [llms.txt]() Data updated: 2026-04-01T08:45:59.815Z Evaluation of general language understanding capabilities on a specific subset of the GLUE test set tasks, reporting scores for MNLI, SST-2, CoLA, QNLI, and MRPC. 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) | GLUE (test) | | Metric | MNLI Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 6 of 6 | | 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 - [MNLI Accuracy](): Higher is better; unit: PERCENT. - [SST-2 Accuracy](): Higher is better; unit: PERCENT. - [CoLA Accuracy](): Higher is better; unit: PERCENT. - [QNLI Accuracy](): Higher is better; unit: PERCENT. - [MRPC Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | MNLI Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Full | 86.89 | Not recorded | 69cca6a20636e5bf4afcbc50 | | rsLoRA | 85.54 | Not recorded | 69cca6a20636e5bf4afcbc50 | | CG-LoRA (no shift) | 85.51 | Not recorded | 69cca6a20636e5bf4afcbc50 | | CG-LoRA (shift) | 85.35 | Not recorded | 69cca6a20636e5bf4afcbc50 | | LoRA-One | 85.03 | Not recorded | 69cca6a20636e5bf4afcbc50 | | LoRA-GA | 83.02 | Not recorded | 69cca6a20636e5bf4afcbc50 | ### Full MNLI Accuracy: 86.89. Result ID: 69ccdb3447f34ce924b6a103. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"RoBERTa-base","Evaluation Protocol":"finetuned"}. This record does not assert independent reproduction. ### rsLoRA MNLI Accuracy: 85.54. Result ID: 69ccdb3447f34ce924b6a104. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"RoBERTa-base","Evaluation Protocol":"finetuned","LoRA Rank":"8"}. This record does not assert independent reproduction. ### CG-LoRA (no shift) MNLI Accuracy: 85.51. Result ID: 69ccdb3447f34ce924b6a107. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"RoBERTa-base","Evaluation Protocol":"finetuned","LoRA Rank":"8","Initialization":"shifted from W0 + ηB0A0"}. This record does not assert independent reproduction. ### CG-LoRA (shift) MNLI Accuracy: 85.35. Result ID: 69ccdb3447f34ce924b6a108. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"RoBERTa-base","Evaluation Protocol":"finetuned","LoRA Rank":"8","Initialization":"starting from W0"}. This record does not assert independent reproduction. ### LoRA-One MNLI Accuracy: 85.03. Result ID: 69ccdb3447f34ce924b6a106. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"RoBERTa-base","Evaluation Protocol":"finetuned","LoRA Rank":"8"}. This record does not assert independent reproduction. ### LoRA-GA MNLI Accuracy: 83.02. Result ID: 69ccdb3447f34ce924b6a105. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-base","Evaluation Protocol":"finetuned","LoRA Rank":"8"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.