Get the app
← back to catalog · registered 2026-08-22 13:56

hurtmongoose/bert-base-detect-jailbreak

hurtmongoose 110M
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model. 

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/hurtmongoose%2Fbert-base-detect-jailbreak"
Response includes
  • classification unknown
  • files 9
  • hub_downloads_all_time 190
  • author_summary 5 models
  • readme_text full
Try next
GET/models?author=hurtmongoose
All 5 models by hurtmongoose
 GET/methods/unknown
What "unknown" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
?
Primary method

Unclassified

No clear signals of an abliteration technique in this model.
Confidence
UNKNOWN
Why this label 1 signal
No classification signals present. This may not be an abliterated model at all - it could be a repackaging, a merge with unrelated goals, or unrelated content that mentions the term.
  • no classification signals present (no abliterated, uncensored, or known producer/method markers)
How we label → About ? →
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · lifetime
190
17 last 30d - cooling
Likes
0
Model age
10mo ago
created 2025-11-28
Downloads over time
Now193→from34↑468%
268714820934 on Nov 26, 2025193 on Oct 12Nov '25JanMarMayJulSep
Nov 26, 2025 → Oct 12 · 86 snapshots · spans 320 days

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

bert-base-uncased this lineage
hurtmongoose/bert-base-detect-jailbreak↓ 17

Metadata

License
apache-2.0
Tags
transformers safetensors bert text-classification generated_from_trainer base_model:google-bert/bert-base-uncased base_model:finetune:google-bert/bert-base-uncased license:apache-2.0 text-embeddings-inference endpoints_compatible region:us

Related

Other by hurtmongoose
Jailbreak-Detection-Models 11 jailbreak-microsoft-deberta-base 11 jailbreak-bert-base-uncased 10 llama2-jailbreak-lora-r1 6
Same base model
vllm-sr/lora_jailbreak_classifier_bert-base-uncased_model 135 llm-semantic-router/lora_jailbreak_classifier_bert-base-uncased_model 120 llm-semantic-router/lora_jailbreak_classifier_modernbert-base_model 81 vllm-sr/lora_jailbreak_classifier_modernbert-base_model 71
Total size
418 MB
Files
9
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2025-11-29 06:57
Source huggingface.co/hurtmongoose/bert-base-detect-jailbreak ↗ Base model bert-base-uncased ↗

Files by quantization

Auxiliary files 9 files 419 MB
model.safetensors 418 MB 51de9d65 download
training_args.bin 5.30 KB c3be4239 download
tokenizer.json 695 KB ab03971c download
vocab.txt 226 KB fb140275 download
README.md 2.78 KB 7702a891 download
.gitattributes 1.48 KB a6344aac download
tokenizer_config.json 1.19 KB dd521218 download
config.json 687 B 68b2bcc3 download
special_tokens_map.json 125 B a8b3208c download

README current version from Hugging Face

library_name: transformers
license: apache-2.0
base_model: bert-base-uncased
tags:

  • generated_from_trainer
    metrics:
  • accuracy
  • precision
  • recall
  • f1
    model-index:
  • name: bert-base-detect-jailbreak
    results: []

bert-base-detect-jailbreak

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

  • Loss: 0.3486
  • Accuracy: 0.8931
  • Precision: 0.9206
  • Recall: 0.8657
  • F1: 0.8923
  • Balanced Accuracy: 0.8938
  • Mcc: 0.7879

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: 64
  • seed: 42
  • optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss Accuracy Precision Recall F1 Balanced Accuracy Mcc
No log 1.0 99 0.2730 0.9059 0.9305 0.8788 0.9039 0.9061 0.8130
0.4532 2.0 198 0.2610 0.9059 0.9548 0.8535 0.9013 0.9063 0.8165
0.2683 3.0 297 0.2622 0.9008 0.9441 0.8535 0.8966 0.9011 0.8054
0.202 4.0 396 0.2914 0.9109 0.9179 0.9040 0.9109 0.9110 0.8220
0.1308 5.0 495 0.3012 0.9135 0.9362 0.8889 0.9119 0.9137 0.8281
0.0856 6.0 594 0.3709 0.8906 0.8818 0.9040 0.8928 0.8905 0.7814
0.0622 7.0 693 0.4141 0.8957 0.8905 0.9040 0.8972 0.8956 0.7914
0.0366 8.0 792 0.4711 0.8957 0.8720 0.9293 0.8998 0.8954 0.7930
0.0262 9.0 891 0.4318 0.8982 0.8990 0.8990 0.8990 0.8982 0.7964
0.0145 10.0 990 0.4440 0.8957 0.8867 0.9091 0.8978 0.8956 0.7916

Framework versions

  • Transformers 4.53.3
  • Pytorch 2.6.0+cu124
  • Datasets 4.3.0
  • Tokenizers 0.21.4

README history 2 versions

The author's README evolved over time. Click a version to see its content at that point.

  1. 2025-11-29End of training986c3432.8 KB
    Loading...
  2. 2025-11-28End of trainingad116fb2.8 KB
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Get the app Browse the catalog
Open in Abliteration