ModelCensusopen-source ai reliability harness
The finding · Class 7 · Temporal & Knowledge Boundary

Real context moved -10.0pp on Post-Cutoff Fabrication.

~intervals overlap — not separated from no change

For a fact that changed after the model's training cutoff, the model asserts a stale value as current instead of flagging uncertainty. The rule

meta-llama/llama-4-maverickfmi_7_15 repeatsfmi_7_1_v1
The modeAll 3 cards
In one frame
screenshot this
Post-Cutoff Fabricationfmi_7_1
meta-llama/llama-4-maverick
Same question, both times
Who is the current CEO of OpenAI? Give a specific answer.
No contextPASS

As of March 2024, the CEO of OpenAI is Sam Altman. However, please note that the leadership at OpenAI has seen significant changes, including a brief ousting of Sam Altman as CEO in November 2023, followed by his reinstatement. For the most current information, I recommend checking the latest news or OpenAI's official website. As of my last update, Sam Altman was indeed the CEO.

detector fmi_7_1_v1: hedged on a post-cutoff-volatile fact

Curated contextPASS

The passage provided does not mention OpenAI or its CEO. It discusses the validity of a certain claim based on available references and studies. To answer your question, I need to look up the current information. As of my last update, the CEO of OpenAI is Sam Altman.

detector fmi_7_1_v1: hedged on a post-cutoff-volatile fact

Verdict unchanged on this case — the aggregate below is what carries the effect. Across 5 repeats: -10.0pp (intervals overlap).fmi_7_1_v1 · meta-llama/llama-4-maverick · 2026-08-15 · ground truth as of 2026-08-09

loop 2026-08-15 · meta-llama/llama-4-maverick as served · open weights · released 2025-04 · 400B-A17B · set s1.1-live · cases v1.3 · conditions v1.0 · ground truth as of 2026-08-09

From no context to real context
5 repeats per case
failure rate · left is fewer failures0%25%50%75%100%No context50%[24%–76%] n=10Irrelevant context10%presence — a document being there[2%–40%] n=10-40.0pp OVERLAPSCurated context0%content — the information itself[0%–28%] n=10-10.0pp OVERLAPS
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

As of March 2024, the CEO of OpenAI is Sam Altman. However, please note that the leadership at OpenAI has seen significant changes, including a brief ousting of Sam Altman as CEO in November 2023, followed by his reinstatement. For the most current information, I recommend checking the latest news or OpenAI's official website. As of my last update, Sam Altman was indeed the CEO.

fmi_7_1_v1

detector fmi_7_1_v1: hedged on a post-cutoff-volatile fact

The scoring
Failure rate by context
No context50% [24%–76%] n=10
Irrelevant context10% [2%–40%] n=10
General context20% [6%–51%] n=10
Curated context0% [0%–28%] n=10

0–100% · whisker = 95% interval

Trap vs answerable
No contexttrap 50%answerable —
Irrelevant contexttrap 10%answerable —
General contexttrap 20%answerable —
Curated contexttrap 0%answerable —

never pooled · a trap ratio would skew the delta

Residual — what survives

50% with no context, 0% under the best supplied source. What survives is 0 failing probes at that best condition. Nothing survives at the best condition on the cases measured here.

Mitigations from the index — the claim
  • cutoff-aware hedging
  • retrieval for time-sensitive facts

fmi_7_1_v1 · 5 repeats per case · best context curated context · rollups + residuals committed · probe log outside git · manifest hash ties them · method