Repository navigation
Add a test on private writing that keeps the texts out of the repository - #5
Merged
Merged
Conversation
…ut of the repository private_eval scores documents listed in RAW_DATA_DIR/private with Clef, Binoculars and both classifiers, then reports how many each detector flags using the bar and combination weights already set on the long set's calibration half. The detectors' entry points now accept any list of texts.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
Every human text in the study is public, so it may be in a model's training data, and recent public text turned out to include AI-assisted writing. Private documents written before language models were common are the clean test of false accusations.
What this adds
detecteval.private_evalwith three steps:clef,localandreport. It readsRAW_DATA_DIR/private/manifest.jsonl, scores each document at 75, 150, 300 and 500 words, and reports how many each detector and combination flags.RAW_DATA_DIR/private. The report holds counts and rates only.binoculars.run,classify.runandscore_long.scorenow accept any list of texts.Validation
bash ci/run.sh: ruff clean, 28 tests pass (one new test for the private-text loader).Not included
No private documents have been scored yet; this PR adds the tooling only.