🔬 DetectorScope 🔬

AI text detectors that show their work.

DetectorScope turns a black-box AI text detector into one whose every prediction is a sum of named, validated features. It can tell you which properties of a text made it predict "AI" and where in the text they are located.

Try our demo! 🤗 https://huggingface.co/spaces/liamdugan/detector-scope-demo

pip install detectorscope

from detectorscope import AuditableDetector

# Load the AI detector
detector = AuditableDetector.from_pretrained(
  "liamdugan/detector-scope"
)

# Run the AI detector on some input
output = detector(
  "Like I said, this is about accuracy, about speed, and about cost."
)

# Print the results
print(output)

Detection(label='ai', p_ai=0.88, features=[
  Feature(importance=+0.23, description='the word about used as a preposition'),
  Feature(importance=+0.14, description='repeated phrases or sequences in text'),
  Feature(importance=-0.05, description='conversational filler phrase like i said'),
  ...])

Using DetectorScope you can:

• 🔍 Query a strong AI detector that is transparent about the features it uses
• 🖱️ Hover over the text to see the specific "AI Phrases" the detector is using
• 🔁 Convert your own black-box detector into an auditable one by using our pipeline
• 🎓 Analyze LLM Outputs so that you understand which models have which tendencies

News

• [25 Oct 2026] DetectorScope will appear at EMNLP 2026 in Budapest, Hungary!

Overview

_{The pipeline: SAEs on the detector's residual stream, a description and validation score per feature, a threshold on that score, and a sparse head over the features that pass.}

A DetectorScope detector is a concept bottleneck over interpretable sparse-autoencoder features: the language model's residual stream is encoded by SAEs, each feature gets a natural-language description that is validated against its own activations, and a sparse linear head over the validated features makes the decision. The prediction is therefore a sum of per-feature contributions, and every contribution has a description and a location in the text.

Results

_{F1 (COLING) and TPR at 5% FPR (RAID) against the interpretability threshold, in domain (grey) and out of domain (white). FT-Model+SAE matches the underlying neural detector even at thresholds of 0.7 and above.}

Across both benchmarks applying our auditable pipeline has minimal adverse effects on performance. The auditable detectors stay close to the fine-tuned black-box classifier they are built on. The full tables are in the paper.

Installation

From PyPI, in a fresh environment:

python -m venv .venv && source .venv/bin/activate
pip install detectorscope

From source, with the extras for the pipeline, the experiments and the demo:

git clone https://github.com/google-deepmind/detectorscope && cd detectorscope
pip install -e ".[all]"

Usage

Any released detector loads by its Hub id; the other settings of a repo are its branches.

from detectorscope import AuditableDetector

detector = AuditableDetector.from_pretrained("liamdugan/detector-scope-peer-review-gpt4o")
detector = AuditableDetector.from_pretrained("liamdugan/detector-scope", revision="base-sae")

detector(text) returns a Detection with label, p_ai, logit, bias and features, every feature active in the text with a non-zero effect, largest first. importance is how much the logit falls if the feature is silenced, and logit == bias + sum(importances). A list of texts returns a list.

Set return_highlights=True to see where each feature is localized within the text:

output = detector(text, return_highlights=True)
output.tokens                   # the text as tokens
output.features[0].highlights   # [(token index, activation), ...]

You can use the tau parameter to control the tradeoff between interpretability and accuracy. Every feature has an auto-interpretation score, how well its description predicts where it fires, and tau is the score a feature must reach before the head is allowed to use it. A higher tau gives the head fewer, better-described features, so every item of evidence it reports is one you can read and check, at some cost in accuracy. A lower tau admits more features, including ones whose descriptions are loose, and the detector gets more accurate and less legible. The default is 0.85.

detector = AuditableDetector.from_pretrained("liamdugan/detector-scope", tau=0.95)

tau runs from 0.00 (every feature) to 0.95 in steps of 0.05, and each branch carries a head at every step, so changing it downloads nothing. It can also be set per call, which is the easiest way to compare the same text at several settings:

output = detector(text, tau=0.50)   # more features, higher accuracy
output = detector(text, tau=0.95)   # the few best-described features
detector.heads                      # every (head, tau) pair the repo holds

head="xgboost", on from_pretrained or on the call, selects the gradient-boosted head at the same tau instead of the logistic-regression one.

Models

We provide the following model collections for download via Huggingface:

Each repo holds the paper's three settings as branches. ft-model-sae fine-tunes Gemma-2-2b as a classifier and fine-tunes the SAEs on it; ft-sae keeps the unmodified Gemma-2-2b and fine-tunes only the SAEs; base-sae keeps both the unmodified Gemma-2-2b and the unmodified Gemma Scope SAEs. main is the setting that performed best on each domain: ft-sae for the open domain, ft-model-sae for peer reviews. Every branch bundles its language model, so nothing else is downloaded and no licence gate is involved.

The defaults are set by hand: the logistic-regression head over features whose auto-interpretation score is at least 0.85, with the decision threshold fixed on the development split. Each branch also carries the XGBoost head and both heads at every other threshold.

Reproducing the paper

The pipeline that builds a detector is the package's stages, run as python -m detectorscope.<stage>: finetune (the classifier), train_sae (SAEs on its residual stream), data.extract and data.process (per-token activations, per-document caches, log-odds ratios, feature statistics), autointerpret (descriptions and their validation, through Gemini; set GEMINI_API_KEY), model.detectors.heads.training.train_logreg / train_xgboost (the heads) and evaluate. Stages 0 to 2 need a GPU; the rest runs on CPUs.

The experiments are four scripts under scripts/: hparam_sweep.py (the SGD + L1 grid search), counterfactual_ablation.py (zero-ablation vs. targeted paraphrase of single features), paraphrase_ablation.py (the feature-guided paraphrase attack) and regex_recall_ablation.py (regex recall of the descriptions). Every figure and table of the paper comes from scripts/visualization/analysis.ipynb, and --help on any script lists its arguments.

demo/ is the Gradio app, which runs locally with python demo/app.py and as the Hugging Face Space; scripts/huggingface_hub/ exports the detectors to the Hub and publishes the Space.

Citation

@misc{dugan2026detectorscope,
  title={Building Auditable AI Text Detectors with Sparse Autoencoders},
  author={Liam Dugan and Callum McDougall and Matthew Denton and Phillip Pham and Christine Kaeser-Chen and Chris Callison-Burch},
  year={2026},
  eprint={2610.30006},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2610.30006},
}

Acknowledgements

This research was developed with funding from the Defense Advanced Research Projects Agency's (DARPA) SciFy program (Agreement No. HR00112520300) and is based upon work supported in part by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA), via 56000026C0019 (the BENGAL program). The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of DARPA, ODNI, IARPA, the Department of Defense, or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright annotation therein.

Disclaimer

The detectors are research artifacts released to support the paper. They make mistakes, including on text written by people, and their accuracy outside the domains and generators they were trained on is not established. They must not be used as the basis for decisions about individuals, such as academic integrity, hiring or publication decisions, and the authors and Google accept no liability for any use of them. This is not an official Google product.

License

Copyright 2026 Google LLC. The software is licensed under the Apache License, Version 2.0; all other materials under the Creative Commons Attribution 4.0 International License (CC-BY). Both are distributed on an "AS IS" basis, without warranties or conditions of any kind; see the licenses for the governing language. Adapted third-party code carries its licence in the header of the module that contains it.