Index of every page on this site: llms.txt
Avatar for Ruoyu Zhao (Roi)

Ruoyu Zhao (Roi)

THU EE+Stats -> CityUHK CS. ML, CV, Optim, Brain-Inspired Intelligence. PhDing; Startup; Fitness; Sleeping.

2 skills 149 Hong Kong Synced

thuroi0787/absent-author

Research and writing can be automated. Authorship cannot. Evidence list + agent skills for spotting AI-produced papers no human stood behind.

2 skills 149
npx skilld add thuroi0787/absent-author --all
  • Final "author pass" for an AI-assisted ML research paper — refine from high level (story, research taste, scale vs. claims, pivots) through argument structure and verification down to sentence-level sloppy-AI traces (em-dash floods, "not X but Y", defensive writing, coined terms, performative honesty), so a responsible human author is visibly present. Use when polishing, de-slopping, or revising a paper draft before submission together with its human author. If asked to make AI text "read human" or pass review or detectors without a human author, it runs audit-only and returns questions for the author. Not an AI-detector evasion tool.

  • Screen an ML research paper (as reviewer, AC, or co-author) for evidence that it was produced by AI without a responsible human author — unsteered research (failed method pivoted into an "audit" paper, theoryslop, toy scale with grand claims, phantom experiments, text–table mismatches), unchecked artifacts (hallucinated references, chatbot residue, watermarks), and unpolished AI-led writing (em-dash floods, defensive writing, coined terms, performative honesty). Produces an evidence-cited grade on three axes — writing W0–W3, research steering R0–R3, ordinary quality Q — plus integrity flags and an auditability rating; never an "AI probability". Use when asked whether a paper is AI slop, to triage a review pile, or to check for an absent author before reviewing.

Curated skills from @thuroi0787. Checked GitHub just now