loop 2026-08-15 · openai/gpt-5-mini as served · closed weights · released 2025-08 · set s1.3-live · cases v1.6 · conditions v1.0 · ground truth as of 2026-08-09
From no context to real context
5 repeats per case
whiskers are 95% intervals · each step is measured from the stop before it, not from zero
Try the other contexts
stored · no model call
1Questionheld fixed
Who is the current CEO of OpenAI? Give a specific answer.
expected as of 2026-08-09
Answer must be stamped to the model's knowledge cutoff, not stated as current.
2Contextchange me
Nothing supplied. The baseline.
3ResponseNo context
PASS
Sam Altman. (That was accurate as of my last update in June 2024 — I can look up the latest confirmation if you’d like.)
fmi_7_1_v1
detector fmi_7_1_v1: hedged on a post-cutoff-volatile fact
The scoring
Failure rate by context
No context30%[8%–69%] n=20 · eff 5
Irrelevant context25%[6%–65%] n=20 · eff 5
General context10%[1%–52%] n=20 · eff 5
Curated context30%[8%–69%] n=20 · eff 5
0–80% · whisker = 95% interval
Trap vs answerable
No contexttrap 30%answerable —
Irrelevant contexttrap 25%answerable —
General contexttrap 10%answerable —
Curated contexttrap 30%answerable —
never pooled · a trap ratio would skew the delta
Residual — what survives
35% with no context, 10% under the best supplied source. What survives is 2 failing probes at that best condition, the detector's reason on most of them being: asserted a volatile post-cutoff fact as current without hedging.
Mitigations from the index — the claim
—cutoff-aware hedging
—retrieval for time-sensitive facts
fmi_7_1_v1 · 5 repeats per case · best context general context · rollups + residuals committed · probe log outside git · manifest hash ties them · method