Slop Detector
A simple model to detect AI-written content.
Model Details
This model is a PEFT of bert-base-cased.
It was trained on a dataset of scraped paragraphs from Wikipedia, as well as AI rewrites of said paragraphs.
More detail about the training data can be found here.
Usage
The basic usage is as follows:
from transformers import pipeline
classifier = pipeline("text-classification", model="gouwsxander/slop-detector-bert")
inputs: str | list[str] | dict[str, str] | list[dict[str, str]] = ...
results = classifier(inputs)
Texts with label LABEL_1 are estimated to be produced by an AI.
You may get a warning about weights being initialized, but these can be ignored.
Limitations
- 512 token context window
- Trained one paragraph at a time
- Only trained on Wikipedia-style text
- Only trained on AI data generated by GPT 5 Nano
- Downloads last month
- 121
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 1 Ask for provider support
Model tree for gouwsxander/slop-detector-bert
Base model
google-bert/bert-base-cased
from peft import PeftModel from transformers import AutoModelForSequenceClassification base_model = AutoModelForSequenceClassification.from_pretrained("bert-base-cased") model = PeftModel.from_pretrained(base_model, "gouwsxander/slop-detector-bert")