Diff

baseline p4-29-reply-30-days · fafd4aa5a67e against candidate p4-30-reply-14-days · 8409b23e21c1

Both runs' records agree. Every check behind this is on Proof

The short answer

The runs first differ at draft, where its output changed.

Every stage before it is the same. Both stages after it differ. A recorded setting also changed later: refund_days at policy_check, 30 → 14.

questionSame
draftFirst difference
policy_checkChanged after it
sendChanged after it

Checked from the records. The same answer every time.

What changed between the two runs

The two runs differ.

First difference: stage 2 · draft.

Stages identical before it: 1 of 1.

Both runs' records agree. question was the same in both. No recorded setting of draft differs. Its output differs, and all 2 stages after it changed too.

no recorded cause: nothing recorded differs at 'draft' (same instrument, version and settings), and its instrument declares itself not rederivable, so its output can vary on its own; the next recorded change is at 'policy_check': refund_days differs

onetrace diff runs/p4-29-reply-30-days runs/p4-30-reply-14-days --text

recorded as a straight line (no graph recorded)

questionSame
draftFirst difference
policy_checkChanged after itFirst recorded changerefund_days 30 → 14
sendChanged after it

A link shows which recorded output a stage read. It does not by itself prove that one stage caused the next.

Settings that changed

Stage	Setting	Baseline	Candidate
policy_check	refund_days	30	14

The first difference is at stage 2 · draft. No smaller set was reported.

The text diff

Shown with at most 200 lines per output (the SDK's default).

Open the text diff on its own page

first difference: draft — output text changed; settings unchanged

draft · First difference

settings: none changed

draft.json application/json Baseline: verified Candidate: verified

1 line removed · 1 added

11  {
22    "question": "I bought a kettle last week and it leaks. Can I return it for a refund?",
3-   "reply": "We're sorry the kettle leaks. Please send it back within the refund window for a full refund."
3+   "reply": "Thanks for letting us know. A leaking kettle can be returned within the refund window, and we'll refund you in full."
44  }

policy_check · Changed after it

settings: 1 changed

refund_days: 30 → 14

input draft.json (operator-authored) changed: 28ee91f1478b -> 65f85c1e08b0; content not stored

checked.json application/json Baseline: verified Candidate: verified

1 line removed · 1 added

11  {
22    "question": "I bought a kettle last week and it leaks. Can I return it for a refund?",
3-   "reply": "We're sorry the kettle leaks. Please send it back within 30 days of delivery for a full refund."
3+   "reply": "Thanks for letting us know. A leaking kettle can be returned within 14 days of delivery, and we'll refund you in full."
44  }

send · Changed after it

settings: none changed

input checked.json (operator-authored) changed: 4f4a22093f01 -> e7265ef350ac; content not stored

sent.json application/json Baseline: verified Candidate: verified

1 line removed · 1 added

11  {
2-   "reply": "We're sorry the kettle leaks. Please send it back within 30 days of delivery for a full refund."
2+   "reply": "Thanks for letting us know. A leaking kettle can be returned within 14 days of delivery, and we'll refund you in full."
33  }

Command: onetrace diff runs/p4-29-reply-30-days runs/p4-30-reply-14-days --text

Repeat this yourself:

onetrace diff runs/p4-29-reply-30-days runs/p4-30-reply-14-days

Other runs that recorded the same value

No recorded settings, code or input digest at stage 2 · draft differs between the two runs, so there is no changed value to look for in other runs.

What may have happened

This demo kept no explanation of this comparison.

This record shows what was verified and evaluated at the time — not what was ultimately true.