ICLR 2027 · reviews out Nov 5
Draft your ICLR 2027 rebuttal from your OpenReview reviews
Upload your paper, bring in your reviews with one click from OpenReview, and get a reply to every concern that points to the passage in your paper, ranked by which ones can move your score. You edit and post; nothing goes to OpenReview without you.
Importing reviews is free. 3 free credits at sign-up cover a review of your paper and one rebuttal draft.
What you get
Every concern, split out
Each review is broken into its weaknesses, questions and requests, with the reviewer's rating and confidence kept alongside, so nothing gets a vague one-line answer.
Ranked by what can move the score
Each concern is marked as one that can move the score, one that needs a clarification, or a minor point. Spend the two weeks on the first group.
A draft reply that cites your paper
Written as the authors, ready to edit, pointing to the section and page that answers the concern. Results you haven't run yet are left as placeholders, never invented.
A plan for the discussion window
What to run or revise before you post, each item tagged hours or days, plus a general response to all reviewers summarising the changes.
Bring your reviews in three ways
In the Rebuttal tab of your paper. Importing is free; you pay only when you draft the replies.
One click from OpenReview
Drag the Send to reviewpaper button to your bookmarks bar, open your paper's OpenReview forum and click it. It reads only that page.
Paste the page
Select all on the forum, copy and paste. Extra page text is fine; only the reviews are kept.
PDF or screenshots
Print the forum to PDF, or add screenshots from your phone, up to 4 MB in total.
Reviews not out yet? Start before Nov 5.
Run three simulated ICLR reviews on the version you submitted, then choose Use predicted reviews in the Rebuttal tab: you get draft answers to the likely objections and a list of experiments to start now, so they are running when the discussion opens. When the real reviews land, the draft shows how many of their concerns were predicted.
ICLR 2027 rebuttal dates
| Reviews released, author–reviewer discussion opens | Nov 5, 2026 |
|---|---|
| Author–reviewer discussion ends (last PDF revision) | Nov 18, 2026 |
| Reviewer–AC discussion and meta-reviews | Nov 19 – Dec 16, 2026 |
| Final decisions | Dec 16, 2026 |
From the official ICLR 2027 dates page, which gives no time zone for the discussion. Treat Nov 18 as end of day Anywhere on Earth, and post early: a reply on Nov 6 can get a follow-up, one on Nov 17 usually can't.
Why the ICLR discussion is worth it
- Scores move. At ICLR 2025, 55% of papers had at least one reviewer change their rating during the discussion. Borderline papers are decided here.
- It's a conversation. There is no limit on the number of comments, so an early, specific reply can get a follow-up question you can still answer.
- You can revise the PDF until Nov 18. Say exactly what changed and where; reviewers don't have to re-read every version.
- It's public. Write for the area chair and future readers, not only the reviewer.
Got your scores? See how papers like yours fared with the ICLR acceptance odds calculator.
Two weeks, three reviewers, every concern answered
Upload the paper you submitted to ICLR, import your reviews, and start from a draft instead of a blank OpenReview box.
Importing free · drafting 2 credits
Questions
When can I use it for ICLR 2027?
As soon as your reviews appear on OpenReview: ICLR 2027 releases reviews on Nov 5, 2026, and the author–reviewer discussion closes on Nov 18. Before Nov 5 you can prepare from simulated ICLR reviews of your paper and get draft answers to the concerns most likely to come up.
Does it post to OpenReview for me?
No. You copy the replies (one click copies the general response and every reply, grouped by reviewer), edit them, and post them yourself. ICLR comments are public, so read every line before it goes up.
Will it make up experiments or numbers?
No. Replies only point to what is in your paper. Where an answer needs a new result, the draft leaves a placeholder and the plan lists the experiment to run, so you fill in real numbers.
Is using it allowed under ICLR's AI policy?
ICLR 2027's author AI policy asks authors to disclose how they used LLMs in the paper and holds them responsible for every claim: a falsehood or misrepresentation produced by an LLM counts as a Code of Ethics violation. Treat the draft as a starting point, check every statement against the paper, and if LLM help leads to substantive changes in the paper itself, update its AI-use section.
What does it cost?
Importing your reviews is free. Drafting the replies costs 2 credits. New accounts get 3 free credits, enough for a Standard ICLR review of your paper (1 credit) and one rebuttal draft. More credits are $5 for 5.
Do rebuttals actually change ICLR scores?
Often. At ICLR 2025, 55% of papers had at least one reviewer change their rating during the discussion, and 25% of individual ratings changed, mostly upward. ICLR 2026 was an exception: after an OpenReview security incident, scores were reset to their pre-rebuttal values and area chairs weighed the discussion instead.
Does it work for other venues?
Yes. Paste reviews from OpenReview, CMT or EDAS for any venue. For CVPR's one-page rebuttal, see the CVPR rebuttal guide on how to cut the draft to fit.
More for ICLR 2027 authors: rebuttal guide and reply template · acceptance odds by score · the ICLR review form · ICLR 2027 dates · CVPR rebuttal guide
reviewpaper.ai is not affiliated with ICLR or OpenReview.