Natural Language Inference on GLUE (val)
0.867QNLIGLM-130B
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GLM-130BZero-shot=true, Number of parameters=130B, MIP training=Included other NLI datasets2022.10 | 0.867 | 0.857 | 0.846 | 0.676 | 0.576 | 0.873 | |
| OPT 175BZero-shot=true, Number of parameters=175B, MIP training=false2022.10 | 0.554 | 0.36 | 0.36 | 0.535 | 0.444 | 0.446 | |
| BLOOM 176BZero-shot=true, Number of parameters=176B, MIP training=false2022.10 | 0.509 | 0.355 | 0.355 | 0.577 | 0.39 | 0.316 |