Neutral is an evaluation tool. It rewrites a prompt to remove signals about who is asking, sends the rewritten prompt to a large language model of your choosing, and shows you the answer with the names put back, alongside the answer exactly as the model wrote it.
It is operated by Zachary R. Wang.
Do not use anything this produces as the basis of an employment decision - hiring, firing, promotion, pay, discipline, or any other decision affecting someone's job.
This is stated on every page for a reason. The tool is unproven: at the time of writing no evaluation has established that it reduces bias, and there is published evidence in the project's own repository that its effect may be small. Treating its output as a fair assessment of a real person would be unsafe and, depending on where you are, may be unlawful.
Nothing here is legal, HR, employment, medical, or financial advice. Large language models state incorrect things confidently. Check anything you intend to rely on.
By using this you confirm that you have the right to put the text you enter into it, and that you will not enter:
You are responsible for what you do with the output.
This is provided as is, with no warranty of any kind. It may be wrong, unavailable, or may change without notice. To the fullest extent the law allows, Zachary R. Wang accepts no liability for any loss arising from its use.
The rewriting can fail in ways that are not obvious. It may remove something that mattered, keep something that did not, or attribute part of an answer to the wrong person. The unmodified answer is always shown alongside so you can check.
See the privacy page. This is the most important thing on either of them.
These terms may change. Continuing to use the tool means accepting the current version.
These terms are governed by the law of the State of California, in the United States, and any dispute arising from them belongs to the courts there.
zachary.wang1@sisyphus.website