Slop Blocker local detector 0.3.0 — data attribution

Bundled offline TF-IDF logistic weights were fitted to the eligible grouped
authorship fitting partition of the broad-v2 public-data experiment. Corpus text,
reviews, caches and evaluation datasets are not bundled with the extension.

HC3 (Guo et al.), Reddit ELI5 subset — CC BY-SA 4.0 as stated by dataset card.
https://huggingface.co/datasets/Hello-SimpleAI/HC3
https://arxiv.org/abs/2301.07597
https://github.com/Hello-SimpleAI/chatgpt-comparison-detection#dataset-copyright
https://creativecommons.org/licenses/by-sa/4.0/

SlopIndex (hgaddipati1118), recorded model outputs, commit
80b1b0df7c7ad1f135601deda2aca0553b94bea1 — repository code/statistics MIT;
outputs published as research data; underlying model terms retained.
https://github.com/hgaddipati1118/slop-index

IMDb Large Movie Review Dataset (Maas et al., 2011), fitting portion of official
training reviews — Stanford research corpus; underlying reviews retain rights.
https://ai.stanford.edu/~amaas/data/sentiment/

TweetEval sentiment / SemEval 2017 (Barbieri et al.; Rosenthal et al.) — CC BY 3.0
for sentiment subset, subject to source/platform terms.
https://github.com/cardiffnlp/tweeteval
https://aclanthology.org/S17-2088/

YouTube Spam Collection (Alberto and Lochter, 2015), dated ham-only controls —
CC BY 4.0, underlying source terms retained.
https://archive.ics.uci.edu/dataset/380/youtube+spam+collection

Revisions, byte hashes, partition policy and limitations are recorded in
evaluation/v04-rules/acquisition.json and manifest.json in the source project.
SlopReview and public perceived-slop votes are external checks only, not training
targets. This notice does not assert a unified license for all source material.
No public store release is performed by this local update.

See MODEL_CARD.md and evaluation/v05-lightweight/ADOPTION.md in the source project.
Scores do not prove AI authorship or quality. The user approved local adoption
with known ordinary-writing false positives; Reveal restores any item.
