Zero-Shot Classification
GLiNER2
Safetensors
English
Russian
extractor
safety
pii
ai-security
zero-shot
text-classification
span-categorization
token-classification
guardrails
Instructions to use hivetrace/gliner-guard-uniencoder with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- GLiNER2
How to use hivetrace/gliner-guard-uniencoder with GLiNER2:
from gliner2 import GLiNER2 model = GLiNER2.from_pretrained("hivetrace/gliner-guard-uniencoder") # Extract entities text = "Apple CEO Tim Cook announced iPhone 15 in Cupertino yesterday." result = extractor.extract_entities(text, ["company", "person", "product", "location"]) print(result) - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -20,6 +20,9 @@ tags:
|
|
| 20 |
---
|
| 21 |
# GLiNER Guard — Unified Multitask Guardrail
|
| 22 |
One encoder model that replaces your entire guardrail stack: safety classification, PII detection, adversarial attack detection, intent and tone analysis — all in a single forward pass.
|
|
|
|
|
|
|
|
|
|
| 23 |

|
| 24 |
|
| 25 |
**147M params · GLiNER2 · uniencoder · modernbert multilingual · zero-shot classification, NER and more · no LLM required**
|
|
@@ -275,4 +278,16 @@ Output:
|
|
| 275 |
'intent': 'adversarial',
|
| 276 |
'tone': 'aggressive'}
|
| 277 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 278 |
```
|
|
|
|
| 20 |
---
|
| 21 |
# GLiNER Guard — Unified Multitask Guardrail
|
| 22 |
One encoder model that replaces your entire guardrail stack: safety classification, PII detection, adversarial attack detection, intent and tone analysis — all in a single forward pass.
|
| 23 |
+
|
| 24 |
+
[](https://arxiv.org/abs/2605.05277)
|
| 25 |
+
[](https://huggingface.co/collections/hivetrace/gliner-guard-v1)
|
| 26 |

|
| 27 |
|
| 28 |
**147M params · GLiNER2 · uniencoder · modernbert multilingual · zero-shot classification, NER and more · no LLM required**
|
|
|
|
| 278 |
'intent': 'adversarial',
|
| 279 |
'tone': 'aggressive'}
|
| 280 |
|
| 281 |
+
```
|
| 282 |
+
|
| 283 |
+
```
|
| 284 |
+
@misc{minko2026glinerguardunifiedencoder,
|
| 285 |
+
title={GLiNER Guard: Unified Encoder Family for Production LLM Safety and Privacy},
|
| 286 |
+
author={Bogdan Minko and Sabrina Sadiekh and Evgeniy Kokuykin},
|
| 287 |
+
year={2026},
|
| 288 |
+
eprint={2605.05277},
|
| 289 |
+
archivePrefix={arXiv},
|
| 290 |
+
primaryClass={cs.CR},
|
| 291 |
+
url={https://arxiv.org/abs/2605.05277},
|
| 292 |
+
}
|
| 293 |
```
|