Reading a review honestly before responding
Lesson 4 of 6 · Manuscripts and reviewer responses
1 · Learn the move · Red-team audit
The first instinct on receiving reviews is to draft the defense - which locks in a reading of the reviewer made while stung. The audit comes first, and it runs against you, not them: for each point, the model states the strongest version of what the reviewer means (not the most dismissible), classifies it - right, partly right, or wrong for a stateable reason - and names what the reviewer saw that you did not. A reviewer who misread your figure is telling you the figure is misreadable; that datum survives even when their criticism is technically wrong. Only after the audit does response drafting start, because the audit determines what you concede, and concessions decided early and honestly are what make the eventual defenses credible.
Here are the reviews: [paste]. Before any response drafting: audit them against my paper, not for it. For each numbered point: (1) steelman - the strongest version of what the reviewer means, (2) classification: RIGHT / PARTLY RIGHT / WRONG, with the reason stated, (3) what the reviewer saw that we did not - even when they are wrong. No response language yet. End with the concede list and the defend list.
2 · Your turn. You write the prompt
The reviews landed this morning: one glowing, one bruising, seven points from Reviewer 2 including two that misread Figure 3 and one that stings because it is right. The senior author wants a response plan by Monday and your current draft of point-by-point replies was written angry.
Remember: the AI sees only your prompt, not this page. If the situation isn't in your prompt, it doesn't exist.
Optional. These shape the output when you run your prompt below, not your score.