DerivedFunction1/xlmr-prompt-injection

xlmr-safety-guard

This model is a fine-tuned version of xlm-roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5705
• Binary F1: 0.9512
• Binary Precision: 0.9553
• Binary Recall: 0.9470

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 5691
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Binary F1	Binary Precision	Binary Recall
1.1483	0.5000	14230	1.0458	0.8896	0.8852	0.8942
0.8357	1.0000	28460	0.8305	0.9227	0.9378	0.9080
0.5725	1.5001	42690	0.6750	0.9406	0.9389	0.9422
0.5078	2.0	56918	0.5897	0.9492	0.9537	0.9447

Framework versions

• Transformers 5.6.2
• Pytorch 2.11.0+cu128
• Datasets 4.8.4
• Tokenizers 0.22.2