Add kp-detect-v23 submission

[kprofundis]

Submission

Detector: kp-detect-v23
Author: Kareem Elsamadicy (Independent Researcher)
Contact: kelsamadicy@gmail.com

Ensemble of a transformer-based semantic classifier (DeBERTa-v3-base, 768-D mean-pooled embeddings + logistic regression) and an attack-feature gradient boosting detector (41-D engineered features, 5-seed ensemble, global temperature calibration). Routing by predicted attack type: semantic classifier weighted more heavily for paraphrase attacks, feature-based detector used alone for character-manipulation attacks.

[github-actions]

Eval run succeeded! Link to run: link

Here are the results of the submission(s):

Kareem Elsamadicy (Independent Researcher)

Release date: 2026-05-23

I've committed detailed results of this detector's performance on the test set to this PR.

On the RAID dataset as a whole (aggregated across all generation models, domains, decoding strategies, repetition penalties, and adversarial attacks), it achieved an AUROC of 93.37 and a TPR of 89.76% at FPR=5% and 79.46% at FPR=1%.
Without adversarial attacks, it achieved AUROC of 94.62 and a TPR of 92.14% at FPR=5% and 82.30% at FPR=1%.

If all looks well, a maintainer will come by soon to merge this PR and your entry/entries will appear on the leaderboard. If you need to make any changes, feel free to push new commits to this PR. Thanks for submitting to RAID!