Downloads · 30 days
12
10% of all-time downloads
Mihaiii/Pallas-0.5-LASER-exp2-0.1
Pallas-0.5-LASER-exp2-0.1 is a text generation model from Mihaiii. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as other.
This model has 3 laser interventions applied on Mihaiii/Pallas-0.5 .
Downloads · 30 days
12
10% of all-time downloads
All-time downloads
126
Public
Parameters
34.4B
68.8 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors68.8 GB · 100%
From the Hugging Face model README
This model has 3 laser interventions applied on Mihaiii/Pallas-0.5 .
All interventions were made on mlp layers (meaning: ["mlp.gate_proj.weight", "mlp.up_proj.weight", "mlp.down_proj.weight"]) with a rate of 8.
| Intervention number | Layer number | Validation acc (higher is better) | Validation logloss (lower is better) | Test acc (higher is better) | Test logloss (lower is better) |
|---|---|---|---|---|---|
| 0 | - | 55.263 | 1.651 | 59.868 | 1.463 |
| 1 | 56 | 55.263 | 1.548 | 60.526 | 1.363 |
| 2 | 54 | 55.263 | 1.505 | 61.184 | 1.332 |
| 3 | 51 | 55.263 | 1.488 | 60.526 | 1.336 |