Does Dot keep a corrected revenue definition after a rewording?
Dot kept the corrected revenue definition across a repeated question and a reworded question in separate chats. Both excluded cancelled orders and returned $23,469, matching the $23,469.02 reference.
1 of 1 test case passed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
Corrected metric definition survives a rephrasingPassEvidence
The “Revenue definition” note was saved and made live for all_users before a new chat was opened.
Input 1 · First chat: original question: “As of September 2, 2026, what was our total revenue over the last 90 days?”
Input 2 · First chat: source selection: “customer_support_db”
A separate new chat was opened.
Input 3 · New chat: rephrased question: “As of September 2, 2026, how much did we earn in total over the last 90 days?”
Dot’s question about the order source
Should I use the customer-support orders or the order-tracking orders? They are separate revenue sources. I'll use the 90 days ending September 2, 2026, inclusive, and exclude cancelled orders.
Answer to the original question
Total revenue was $23,469 for the 90 days ending September 2, 2026. Uses customer-support orders placed June 5–September 2 inclusive (UTC), excluding cancelled orders.
Answer to the rephrased question
We earned $23,469 in total revenue over the 90 days ending September 2, 2026 (June 5–September 2, inclusive), excluding cancelled orders.
Transcribed from Proof 1 · Screen recording at 1:52
Every message the tool sent, in order (6)
- Dot’s question about the order source
Printed above.
- Answer to the original question
Printed above.
- Query for the original question · not part of the answer
SELECT currency, COUNT(id) AS total_orders, ROUND(SUM(total_amount), 2) AS total_revenue_usd FROM customer_support_db.public.orders WHERE placed_at >= '2026-06-05 00:00:00+00' AND placed_at <= '2026-09-02 23:59:59.999999+00' AND status != 'cancelled' GROUP BY currency ORDER BY currency - Progress on the rephrased question
I need to choose the right source for customer support orders. I'm considering the dates for the last 90 days and trying to include both September 2 and June 5. Okay, September 2 seems to be 89 days ago, which is fine for inclusion with June 5. I want to explicitly ask for June 5 and September 2 to be inclusive while ensuring no cancelled currencies are included.
- Answer to the rephrased question
The final response, printed above.
- Query for the rephrased question · not part of the answer
SELECT currency, ROUND(SUM(total_amount), 2) AS total_revenue_usd, COUNT(id) AS order_count FROM customer_support_db.public.orders WHERE status != 'cancelled' AND placed_at >= '2026-06-05 00:00:00+00' AND placed_at < '2026-09-03 00:00:00+00' GROUP BY currency
| ✓ | Expected: Both wordings must exclude cancelled orders: $23,469.02 across 42 orders, versus $39,806.75 across 57 orders if cancelled orders are counted.Found: Both answers gave $23,469, matching the $23,469.02 reference; both generated queries excluded cancelled orders. |
With the revenue note active, both fresh chats excluded cancelled orders and returned $23,469 against a $23,469.02 reference, without either question restating the rule. This demonstrates retention across wordings, not a change caused by training.
Order-source clarification. In the repeat chat, Dot asked which of two order sources to use. The user's reply named the customer-support source without adding the revenue rule.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | AI Database Agents → | 28 scenarios · 11 tools |
| Capability | Training → | |
| Scenario | The agent used the wrong definition and the user corrects it → | |
| Tool | Dot → |
Global scenario definition → · Global capability definition → · Dot product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.