File size: 538 Bytes
1fef77b a1e6730 72e2b93 |
1 2 3 4 5 6 7 8 9 10 |
---
license: llama2
---
Experiment for DARE(Drop and REscale), most of the delta parameters can be directly set to zeros without affecting the capabilities of SFT LMs and larger models can tolerate a higher proportion of discarded parameters.
| Model | Average | ARC | HellaSwag | MMLU | TruthfulQA | Winogrande | GSM8K | DROP |
| ------ | ------ | ------ | ------ | ------ | ------ | ------ | ------ | ------ |
| bhenrym14/mistral-7b-platypus-fp16 | 56.89 | 63.05 | 84.15 | 64.11 | 45.07 | 78.53 | 17.36 | 45.92 |
|