Contact
LLM Ground is pre-launch — Phase 1 (the first 25 probes and their runs) hasn't published yet. If you're reaching out before then, about the methodology, a probe idea, a correction, or anything else, we'd genuinely like to hear it.
developers@nocodeweb.orgWhat is worth writing about
- A result you think is wrong. Every score here publishes the prompt, the rubric and the raw output that produced it, precisely so that it can be argued with. If a rubric measures the wrong thing, or a score does not follow from the output beneath it, that is the most useful message we can get.
- A probe worth adding. The library is meant to be jobs people actually do, not exam questions. A task that keeps failing for you in production is a better probe than anything invented at a desk.
- A flaw in the methodology. The methodology page publishes every judge prompt verbatim and keeps a standing list of what has not been verified. A scoring system nobody challenges has not been tested.
- Reuse and citation. Every run page is a permanent URL with its own structured data, and the raw output is served from this origin as plain text. Cite it directly; no permission needed.
There is no form, no ticketing system and no autoresponder — mail goes to a person. Since this is a pre-launch project run by a small team, expect a reply in days rather than hours, and expect a real answer rather than an acknowledgement.
For anything about how this site handles data, see the Privacy Policy.