MNLI
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Natural Language Inference | MNLI (matched) | Accuracy91.7 | 110 | |
| Correlation to Model Behavior Differences | MNLI | Accuracy Correlation0.32 | 93 | |
| Natural Language Inference | MNLI | Accuracy (matched)90.8 | 80 | |
| Natural Language Inference | MNLI (mismatched) | Accuracy91 | 68 | |
| Natural Language Inference | MNLI (test) | Accuracy0.898 | 52 | |
| Backdoor Defense | MNLI (test) | Clean Accuracy91.8 | 44 | |
| Natural Language Inference | MNLI (dev) | Acc (m)90.2 | 44 | |
| Natural Language Inference | MNLI | Accuracy86.2 | 36 | |
| Text Classification | MNLI | Accuracy87.45 | 32 | |
| Classification | MNLI (val) | Accuracy84.17 | 32 | |
| Natural Language Inference | MNLI mm | Accuracy90.7 | 30 | |
| Natural Language Inference | MNLI (val) | Accuracy92.13 | 26 | |
| Natural Language Inference | MNLI | MNLI Accuracy87.4 | 23 | |
| Natural Language Inference | MNLI | Accuracy87.98 | 22 | |
| Natural language inference | MNLI GLUE (test) | Accuracy86.56 | 20 | |
| Natural Language Inference | MNLI few-shot zero-shot | Accuracy71.1 | 16 | |
| Natural Language Inference | MNLI Neg | Accuracy75 | 14 | |
| Structural Bias Evaluation | MNLI | Accuracy98.1 | 14 | |
| Natural Language Inference | MNLI | Normalized Accuracy96.2 | 13 | |
| Natural Language Inference | MNLI-m | Accuracy77.2 | 13 | |
| Natural Language Inference | MNLI Unknown Bias (in-distribution) | Accuracy84.2 | 13 | |
| Natural Language Inference | MNLI HardSP (challenge) | Accuracy83.2 | 13 | |
| Natural Language Inference | MNLI HardCD (challenge) | Accuracy0.803 | 13 | |
| Natural Language Inference | MNLI Hypothesis-only Bias (in-distribution) | Accuracy84.2 | 13 | |
| Natural Language Inference | MNLI Syntactic Bias (in-distribution) | Accuracy84.3 | 13 |
Showing 25 of 75 rows