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

surrey-nlp/roberta-large-finetuned-abbr-unfiltered-plod

surrey-nlp
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/surrey-nlp%2Froberta-large-finetuned-abbr-unfiltered-plod"
Response includes
  • classification unknown
  • files 11
  • author_summary 3 models
  • readme_text full
Try next
GET/models?author=surrey-nlp
All 3 models by surrey-nlp
 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 · 30-day
6
↑ 30,500% in 90 days
Likes
0
Model age
3.5y ago
created 2023-04-20
Downloads over time
Now612→from2↑30,500%
02244496732 on Jul 24, 2024612 on Oct 11612 on Oct 8Jul '24Nov '24Mar '25Jul '25Nov '25MarJul
Jul 24, 2024 → Oct 11 · 155 snapshots · spans 809 days

Metadata

License
cc-by-sa-4.0
Languages
en
Tags
transformers pytorch roberta token-classification generated_from_trainer en license:cc-by-sa-4.0 endpoints_compatible region:us

Related

Other by surrey-nlp
flair-abbr-pubmed-unfiltered 0 flair-abbr-roberta-pubmed-plos-unfiltered 0
Total size
1.32 GB
Files
11
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2024-05-28 08:47
Source huggingface.co/surrey-nlp/roberta-large-finetuned-abbr-unfiltered-plod ↗

Files by quantization

Auxiliary files 11 files 1.32 GB
pytorch_model.bin 1.32 GB 2c843e32 download
training_args.bin 2.92 KB c8f481d6 download
tokenizer.json 1.29 MB 5fa3c40d download
vocab.json 780 KB 4ebe4bb3 download
merges.txt 446 KB 6636bda4 download
README.md 3.92 KB 84334ec4 download
.gitattributes 1.44 KB c7d9f333 download
config.json 915 B 5cfff902 download
tokenizer_config.json 349 B 68512076 download
special_tokens_map.json 239 B 2ea7ad0e download
.gitignore 13.0 B 0348ea97 download

README current version from Hugging Face

license: cc-by-sa-4.0
tags:

  • generated_from_trainer
    metrics:
  • precision
  • recall
  • f1
  • accuracy
    model-index:
  • name: roberta-large-finetuned-abbr
    results: []
    language:
  • en

roberta-large-finetuned-abbr-unfiltered-plod

This model is a fine-tuned version of roberta-large on the PLODv2 unfiltered dataset.
It is released with our LREC-COLING 2024 publication Using character-level models for efficient abbreviation and long-form detection. It achieves the following results on the test set:

Results on abbreviations:

  • Precision: 0.8916
  • Recall: 0.9152
  • F1: 0.9033

Results on long forms:

  • Precision: 0.8607
  • Recall: 0.9142
  • F1: 0.8867

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: 4
  • eval_batch_size: 4
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 6

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.167 0.25 7000 0.1616 0.9484 0.9366 0.9424 0.9376
0.1673 0.49 14000 0.1459 0.9504 0.9370 0.9437 0.9389
0.1472 0.74 21000 0.1560 0.9531 0.9373 0.9451 0.9398
0.1519 0.98 28000 0.1434 0.9551 0.9382 0.9466 0.9415
0.1388 1.23 35000 0.1472 0.9516 0.9374 0.9444 0.9400
0.1291 1.48 42000 0.1416 0.9557 0.9403 0.9479 0.9431
0.1298 1.72 49000 0.1394 0.9577 0.9459 0.9517 0.9470
0.1269 1.97 56000 0.1401 0.9587 0.9446 0.9516 0.9468
0.1128 2.21 63000 0.1410 0.9568 0.9497 0.9533 0.9486
0.1154 2.46 70000 0.1366 0.9583 0.9495 0.9539 0.9493
0.1138 2.71 77000 0.1413 0.9600 0.9502 0.9551 0.9506
0.1117 2.95 84000 0.1313 0.9605 0.9501 0.9552 0.9508
0.0997 3.2 91000 0.1503 0.9577 0.9527 0.9552 0.9507
0.1008 3.44 98000 0.1360 0.9587 0.9536 0.9561 0.9515
0.0909 3.69 105000 0.1435 0.9619 0.9520 0.9569 0.9525
0.0903 3.93 112000 0.1482 0.9619 0.9522 0.9570 0.9528
0.075 4.18 119000 0.1603 0.9616 0.9546 0.9581 0.9537
0.0804 4.43 126000 0.1512 0.9600 0.9560 0.9580 0.9536
0.0811 4.67 133000 0.1435 0.9628 0.9543 0.9585 0.9540
0.0778 4.92 140000 0.1384 0.9616 0.9566 0.9591 0.9548
0.065 5.16 147000 0.1640 0.9622 0.9567 0.9595 0.9550
0.0607 5.41 154000 0.1755 0.9632 0.9562 0.9597 0.9554
0.0587 5.66 161000 0.1643 0.9622 0.9575 0.9599 0.9555
0.062 5.9 168000 0.1663 0.9628 0.9569 0.9598 0.9556

Framework versions

  • Transformers 4.16.2
  • Pytorch 1.11.0
  • Datasets 2.1.0
  • Tokenizers 0.10.3

README history 7 versions

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

  1. 2024-05-28Update README.mdfdcc6dc3.9 KB
    Loading...
  2. 2024-03-21Update README.mddb06be03.8 KB
    Loading...
  3. 2024-03-08Update README.md2825ee33.7 KB
    Loading...
  4. 2024-03-08Update README.md64d4b233.7 KB
    Loading...
  5. 2024-03-07Update README.mdb972e0c3.7 KB
    Loading...
  6. 2024-03-07Update README.md670a3373.6 KB
    Loading...
  7. 2023-04-21update model card README.md529fa723.6 KB
    Loading...

Discussions 1 thread

  1. 2025-01-25PRAdding `safetensors` variant of this modelopen1 💬#1
    Loading...

Continue on Hugging Face →

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