Continual Learning on NLP Task Sequence (MNLI, BOOLQA, RTE, QQP) (test)
77.03FAPRGD-R
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| RGD-RBackbone=Llama2-7B2024.11 | 77.03 | 2.65 | -1.25 | 51.06 | |
| InsCLBackbone=Llama2-7B2024.11 | 76.73 | 2.78 | -1.96 | 50.25 | |
| InsCLBackbone=Mistral-7B2024.11 | 76.17 | 4.43 | -4.02 | 54.08 | |
| EABackbone=Llama2-7B2024.11 | 76.1 | 3.52 | -2.49 | 50.91 | |
| RGD-RBackbone=Mistral-7B2024.11 | 74.91 | 4.37 | -3.92 | 50.77 | |
| EABackbone=Mistral-7B2024.11 | 72.15 | 7.59 | -6.96 | 51.17 |