Minor model card edits
Browse files
README.md
CHANGED
@@ -13,17 +13,18 @@ tags:
|
|
13 |
- jailbreak
|
14 |
- security
|
15 |
- moderation
|
|
|
16 |
---
|
17 |
|
18 |
# Jailbreak Classifier
|
19 |
|
20 |
-
Classifies prompts as jailbreaks or benign. This is a fine-tune checkpoint of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the [jailbreak-classification](https://huggingface.co/datasets/jackhhao/jailbreak-classification)
|
21 |
|
22 |
## Training Details
|
23 |
|
24 |
### Training Data
|
25 |
|
26 |
-
Fine-tuned on the [jailbreak-classification](https://huggingface.co/datasets/jackhhao/jailbreak-classification)
|
27 |
|
28 |
### Training Procedure
|
29 |
|
@@ -31,7 +32,7 @@ Fine-tuned on the [jailbreak-classification](https://huggingface.co/datasets/jac
|
|
31 |
|
32 |
Fine-tuning hyper-parameters:
|
33 |
- learning_rate = 5e-5
|
34 |
-
-
|
35 |
-
-
|
36 |
- lr_scheduler_type = linear
|
37 |
- num_train_epochs = 5.0
|
|
|
13 |
- jailbreak
|
14 |
- security
|
15 |
- moderation
|
16 |
+
- prompt-injection
|
17 |
---
|
18 |
|
19 |
# Jailbreak Classifier
|
20 |
|
21 |
+
Classifies prompts as jailbreaks or benign. This is a fine-tune checkpoint of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the [jailbreak-classification](https://huggingface.co/datasets/jackhhao/jailbreak-classification) dataset.
|
22 |
|
23 |
## Training Details
|
24 |
|
25 |
### Training Data
|
26 |
|
27 |
+
Fine-tuned on the [jailbreak-classification](https://huggingface.co/datasets/jackhhao/jailbreak-classification) dataset.
|
28 |
|
29 |
### Training Procedure
|
30 |
|
|
|
32 |
|
33 |
Fine-tuning hyper-parameters:
|
34 |
- learning_rate = 5e-5
|
35 |
+
- train_batch_size = 8
|
36 |
+
- eval_batch_size = 8
|
37 |
- lr_scheduler_type = linear
|
38 |
- num_train_epochs = 5.0
|