Reviewer FAQ

Common questions from QCs reviewing annotator work on Project Lizard. Newly claimed tasks have 4 hours from claim for Annotators, Reviewers/QC, and Auditors. Returned tasks have 12 hours from return before being recirculated. Finish and submit within the applicable window; once it expires, the task can be pulled from your queue.

Faqs

Section 2

Edge Cases

Workflow guidance for unusual situations reviewers encounter on Project Lizard.

Audit_II return exception: Client audit / Audit_II failures currently route to Audit_I as otherwise normal tasks. Always check the Audit_II feedback tab. These tasks use a no-send-back flow, not the standard QC return instructions below. Rebut incorrect feedback in the specific prompt's Audit_I feedback, or fix correct feedback yourself and resubmit. See the Reviewer Playbook for the full procedure and unpaid scrapped-task guidance.

Getting paid HAI Timer FAQ

Reviewer FAQ

Faqs

Q1 — When Should QC Send To Hold vs. Skipped vs. Unusable?

Q2 — Guidance on "Trick Questions" Or Prompts With Unnecessary Information?

Q3 — How Broadly Should We Apply Table/Chart/Graph Understanding?

Q4 — Is It OK For Prompts To Force The Model To Infer Abbreviations (e.g., NA = North America)?

Q5 — How Should "Difference" Be Interpreted?

Edge Cases

Q6 — Can I rerun an unchanged prompt after a correct model answer?

Q7 — What Should I Do If I Encounter A Completely Blank Task?

Q8 — How Should I Handle Older Tasks Created Before Playbook Updates?

Q9 — What If I Suspect AI Was Used To Generate A Prompt?

Q10 — What Does FINISH_REASON_LENGTH Mean In The Model Generated Answer?