HoldsHolds on claude-haiku-4-5No measured effect on gpt-5-miniNo measured effect on gemini-3.1-flash-lite
Asking for a rewrite
Asking the model to critique then revise its own answer improves final quality over a single pass.
Ask for a draft, then ask the model to criticise it and rewrite.
Why it matters
A second pass can catch a weak answer the first pass missed.
How to use it
In one prompt, ask for a draft, then a short critique, then a final version. Compare it against what a single pass gives you.
Rubric mean up 0.30 points on claude-haiku-4-5. Inside the margin on gpt-5-mini and gemini-3.1-flash-lite.
What the marks mean
- Holds
- No measured effect
Show the per-model numbers
Holds on claude-haiku-4-5. No measured effect on gpt-5-mini and gemini-3.1-flash-lite.
Circulates in practitioner communities. Tested because it circulates, not because it is endorsed.
This tip is OpenAddict's plain-language read of the measured result. The measurement below is the evidence, and it is what the reading has to answer to.
Ledger idC12-critique-revise
What was tested
This claim circulates in practitioner communities as advice about how to write prompts. That it circulates is an input to what gets tested here. It is a reason to test the claim, and it is not evidence for or against it. The result below is the evidence, and it is the only thing on this page that carries weight.
The comparison is paired. Two prompts differ in one respect, the manipulated variable, and are otherwise identical by construction. Nothing here supports a causal reading beyond that pairing.
Result
| Measure | claude-haiku-4-5 | gpt-5-mini | gemini-3.1-flash-lite |
|---|---|---|---|
| Control arm | 3.600n 30 | 4.144n 30 | 4.200n 30 |
| Treatment arm | 3.900n 30 | 4.167n 30 | 4.133n 30 |
| Delta | +0.300 | +0.022 | -0.067 |
| Interval, 95 percent | 0.084 to 0.516 | -0.167 to 0.211 | -0.236 to 0.103 |
| Orbit | Stable60 of 60 records | In free drift60 of 60 records | In free drift60 of 60 records |
Every cell in this table is rubric graded. The score is a grader model reading the answer against a published rubric, not a deterministic check.
Orbit is assigned by the frozen status_v1 rule. On this scale, rubric mean, 1 to 5, the pass threshold is +0.30 and the failure floor is -0.30, each requiring an interval that excludes zero.
Method for this claim
- Task set
- The same 6 domain questions as C09 through C11. The treatment arm asks for a draft, a critique, and a revision in one response; only the final revision is graded.
- Runs per model per arm
- 5
- Scoring
- Rubric graded, via rubric-quality-v1. A grader model reads each answer against a published rubric.
- Pass criterion as written for the pilot
- Mean graded score for treatment exceeds control by at least 0.5 rubric points on a 5 point scale, with the interval excluding zero.
The published verdict comes from status_v1, not from the pass criterion above. The criterion is recorded because it is what the claim was registered with before the run.
Model versions, as recorded
| Model | Version string returned | Delta on this claim | Interval |
|---|---|---|---|
| claude-haiku-4-5 | claude-haiku-4-5-20251001 | +0.300 | 0.084 to 0.516 |
| gpt-5-mini | gpt-5-mini-2025-08-07 | +0.022 | -0.167 to 0.211 |
| gemini-3.1-flash-lite | gemini-3.1-flash-lite | -0.067 | -0.236 to 0.103 |
Reading across models
Sampling was not held constant across vendors, so comparing one model column against another compares two settings as well as two models.
gpt-5-mini rejected the fixed sampling setting and ran at its own default on all 1040 of its calls. The other models ran at temperature 0.