Real context moved 0.0pp on Method/Execution Split.
~intervals overlap — not separated from no change
The model chooses a correct method but executes it wrongly (or vice versa) — a right-approach/wrong-answer or wrong-approach/right-answer split. The rule
A shop sells 17 boxes holding 24 items each, then 8 items are returned. How many items were sold?
To find the number of items sold, we first need to calculate the total number of items sold before any returns. The shop sells 17 boxes with 24 items in each box. So, the total number of items initially sold is 17 * 24 = 408. After 8 items are returned, the total number of items sold becomes 408 - 8 = 400. Answer: 400
detector fmi_4_1_v1: answer matches oracle
To find the number of items sold, we need to first calculate the total number of items sold before any returns, and then subtract the number of items returned. The total number of items initially sold = number of boxes * items per box = 17 * 24 = 408. After 8 items are returned, the total number of items sold = 408 - 8 = 400. Answer: 400
detector fmi_4_1_v1: answer matches oracle
loop 2026-08-15 · meta-llama/llama-4-maverick as served · open weights · released 2025-04 · 400B-A17B · set s1.3-live · cases v1.6 · conditions v1.0 · ground truth as of 2026-08-09
A shop sells 17 boxes holding 24 items each, then 8 items are returned. How many items were sold?
expected as of 2026-08-09
400. Order matters: multiply before subtracting.
Nothing supplied. The baseline.
To find the number of items sold, we first need to calculate the total number of items sold before any returns. The shop sells 17 boxes with 24 items in each box. So, the total number of items initially sold is 17 * 24 = 408. After 8 items are returned, the total number of items sold becomes 408 - 8 = 400. Answer: 400
fmi_4_1_v1
detector fmi_4_1_v1: answer matches oracle
0–20% · whisker = 95% interval
never pooled · a trap ratio would skew the delta
Unchanged by context: 0% with none, 0% with the strongest supplied source. For a control mode that flat line is the result — it is what licenses reading movement elsewhere in this loop as grounding rather than as a document being present. Nothing survives at the best condition on the cases measured here.
- tool-assisted computation
- verify final against method
fmi_4_1_v1 · 5 repeats per case · best context general context · rollups + residuals committed · probe log outside git · manifest hash ties them · method