Token Classification
Transformers
Safetensors
English
modernbert
security
jailbreak-detection
prompt-injection
tool-calling
llm-safety
mcp
Eval Results (legacy)
Instructions to use rootfs/tool-call-verifier with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use rootfs/tool-call-verifier with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("token-classification", model="rootfs/tool-call-verifier")# Load model directly from transformers import AutoTokenizer, AutoModelForTokenClassification tokenizer = AutoTokenizer.from_pretrained("rootfs/tool-call-verifier") model = AutoModelForTokenClassification.from_pretrained("rootfs/tool-call-verifier", device_map="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "model_name": "answerdotai/ModernBERT-base", | |
| "num_labels": 2, | |
| "label2id": { | |
| "AUTHORIZED": 0, | |
| "UNAUTHORIZED": 1 | |
| }, | |
| "id2label": { | |
| "0": "AUTHORIZED", | |
| "1": "UNAUTHORIZED" | |
| }, | |
| "severity": { | |
| "AUTHORIZED": 0, | |
| "UNAUTHORIZED": 4 | |
| }, | |
| "loss": "CrossEntropyLoss (class-weighted)", | |
| "class_weights": [ | |
| 0.5, | |
| 3.0 | |
| ], | |
| "optimization_target": "unauthorized_f1", | |
| "batch_size": 16, | |
| "epochs": 5, | |
| "learning_rate": 3e-05, | |
| "max_length": 1024 | |
| } |