Real context moved -15.0pp on fmi_3_1, but the intervals overlap.
gpt-4o-mini · A constraint set early in a long conversation is silently dropped as the dialogue continues.
loop 2026-08-09 · gpt-4o-mini as served · set s1.0 · cases v1.2 · conditions v1.0 · ground truth as of 2026-08-09
One question, asked twice — both answers, both verdicts, one frame.
Answer in exactly two sentences. What is a confidence interval?
Yes — I can give you a specific answer. It is often reported that distractor; the constraint is immediately adjacent to the task. This is well established. [FIXTURE · fmi_3_1 · bare]
detector fmi_3_1_v1: signature matched — no distractor; the constraint is immediately adjacent to the task.
According to the passage supplied: No distractor; the constraint is immediately adjacent to the task. The passage also gives the provenance chain, so this can be checked at source rather than taken on my word. [FIXTURE · fmi_3_1 · recipe_b]
detector fmi_3_1_v1: no signature match
Question fixed · change the context · stored recordings, no model call.
Answer everything below in exactly two sentences. [long distractor passage] Now summarise the passage.
expected as of 2026-08-09
The two-sentence constraint set early must survive the intervening length.
Nothing supplied. The baseline.
Yes — I can give you a specific answer. It is often reported that two-sentence constraint set early must survive the intervening length. This is well established. [FIXTURE · fmi_3_1 · bare]
fmi_3_1_v1
detector fmi_3_1_v1: signature matched — the two-sentence constraint set early must survive the intervening length.
chained, not independent · small presence step = the content effect is the information
scale 0–80% · whisker is the 95% interval
never pooled · trap ratio would skew the delta
Barely moves. A supplied document does not restore a constraint set forty turns earlier, and the small drop is within the interval and indistinguishable from the presence effect.
- — constraint re-grounding
- — system-level constraint pinning
Scored by fmi_3_1_v1 over 5 repeats per case · best context here was Curated context. Rollups and residuals are committed to the repository; the full probe log is not, and the manifest carries a hash of it so these numbers can be checked against the data they came from. Method.