syx
commited on
Commit
·
5efe432
1
Parent(s):
db3f13d
minor
Browse files
README.md
CHANGED
|
@@ -63,6 +63,15 @@ outputs = model.generate(**inputs)
|
|
| 63 |
response = tokenizer.decode(outputs[0])
|
| 64 |
```
|
| 65 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 66 |
## Citation
|
| 67 |
|
| 68 |
If you use this model in your research, please cite:
|
|
|
|
| 63 |
response = tokenizer.decode(outputs[0])
|
| 64 |
```
|
| 65 |
|
| 66 |
+
## Benchmarks
|
| 67 |
+
|
| 68 |
+
The model has been evaluated on standard benchmarks to verify its performance:
|
| 69 |
+
|
| 70 |
+
- **MMLU**: 69.19% (5-shot)
|
| 71 |
+
- **IFEval**: 73.2% (Prompt Strict-Accuracy)
|
| 72 |
+
|
| 73 |
+
These results demonstrate that the ReLU modification maintains competitive performance while achieving higher sparsity compared to the original model.
|
| 74 |
+
|
| 75 |
## Citation
|
| 76 |
|
| 77 |
If you use this model in your research, please cite:
|