This task involves the work of sharing the Editing and Machine Learning Teams' plans for running an A/B test of Tone Check.
Learning objectives
- 1. What – if any – concerns/issues about Tone Check (and the A/B test we're planning to help evaluate its impact through) are volunteers holding that we are aligned in thinking need to be addressed before the experiment can proceed?
- 2. What – if any – facets of how we've documented Tone Check on-wiki could be improved to make it easier for volunteers to assess the experiment proposal we are asking them to review?
Proposed Wikis
Announcement contents
- Instructions for how volunteers can try the experience
- Instructions on how volunteers can identify edits that triggered Tone check (mainly through public tags: T389897, T388716, T395166)
- Details about:
- How long we anticipate the test to run (e.g. ecenable=2)
- Who will be included in the test (T389231)
- How we will be evaluating the impact of the intervention
- The model: how it was trained and evaluated
- Instructions on how to provide feedback