Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →brycewang-stanford/acmmm-author-response
Use when drafting the ACM MM (ACM Multimedia) rebuttal — turning multiple anonymous reviews into one focused, anonymous response that fixes factual errors first, adds small confirmatory cross-modal results, concedes calibratedly, and stays inside policy (anonymous, no new external links), written for the area chair who decides.
Passed 20 of 22 test cases with the skill, and 13 of 22 without it.
and 9 more in the full report
8 fixed12 passed either way1 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
Please generate a response to the Area Chair for our ACM MM review addressing R1's factual error about missing loss curves (located in Figure 4). Format the output into the standard 5-part structure: [Factual corrections], [New results], [Concessions], [Contested], and [Policy check].
| Checked | Without skill | With skill |
|---|---|---|
| The output contains all five explicit section headers: [Factual corrections], [New results], [Concessions], [Contested], and [Policy check]. | Passed | Passed |
| The [Policy check] block includes explicit pass or fix evaluations for anonymity, external links, and scope change. | Missed | Passed |
The instructions your model reads when the skill loads.
Use this to write the ACM Multimedia rebuttal. It is optional and anonymous, it must not add new external links, and its real audience is the area chair who will weigh it against the reviews in discussion.
Sort every review point into one of three buckets:
| Bucket | Example | Response |
|---|---|---|
| Factual error | "No ablation of the fusion" (there is one, in §4.3) | Point to it precisely; this is your strongest move |
| Fixable gap | "Missing a late-fusion baseline" | Add the small result if you can run it in time |
| Judgment call | "The delta is incremental" | Reframe the significance briefly, or concede narrowly |
Lead with the factual corrections — they are the points an AC can verify and act on fastest.
Write for a reader who is skimming several rebuttals. Make it navigable:
[R1, R2 shared concern: fusion not shown to matter]
-> Table A (new): leave-one-modality-out; audio-text term = +X of the +Y total gain.
[R1 factual: "no user study"]
-> §5 reports a user study, N raters, agreement kappa; see supplement S3.
[R3 judgment: "incremental over late fusion"]
-> Our gain over the late-fusion baseline is in Table 1; the mechanism (Table A) is the novelty.
[R2 request: clarify synchronization assumption]
-> Conceded; we will state it explicitly in §3 (one sentence).
Group shared concerns once; do not repeat the same answer per reviewer.
test cases passed, vs 13 / 22 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.