ModelCensusopen-source ai reliability harness
The finding · Class 5 · Robustness & Consistency

Real context moved -3.3pp on Paraphrase Non-Invariance.

~intervals overlap — not separated from no change

Meaning-preserving paraphrases of the same question yield materially different answers. The rule

microsoft/phi-4fmi_5_25 repeatsfmi_5_2_v1
The modeAll 10 cards
In one frame
screenshot this
Paraphrase Non-Invariancefmi_5_2
microsoft/phi-4
Same question, both times
How many bones are in the adult human body?
No contextPASS

As a large language model, I cannot be relied upon for definitive information on medical or health-related matters. It is recommended to consult a qualified medical professional for accurate and up-to-date information. In a typical adult human skeleton, there are 206 bones. This number can vary slightly due to anatomical variations, such as extra ribs or additional small bones in the hands or feet.

detector fmi_5_2_v1: answers invariant across paraphrases

General contextPASS

In an adult human, the total number of bones is typically 206. This count is reached after the fusion of certain bones during the developmental process from infancy to adulthood. For example, infants are born with approximately 270 bones, but as they grow, some of these bones fuse together, resulting in the standard 206 bones in a fully developed adult skeleton.

detector fmi_5_2_v1: answers invariant across paraphrases

Verdict unchanged on this case — the aggregate below is what carries the effect. Across 5 repeats: -3.3pp (intervals overlap).fmi_5_2_v1 · microsoft/phi-4 · 2026-08-15 · ground truth as of 2026-08-09

loop 2026-08-15 · microsoft/phi-4 as served · open weights · released 2024-12 · 14B · 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
failure rate · left is fewer failures0%10%20%30%40%No context7%[2%–21%] n=30Irrelevant context3%presence — a document being there[1%–17%] n=30-3.3pp OVERLAPSGeneral context0%content — the information itself[0%–11%] n=30-3.3pp 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
How many bones are in the adult human body?

expected as of 2026-08-09

206. 206 in a typical adult. Widely misremembered, which is what makes the phrasings diverge.

2Contextchange me

Nothing supplied. The baseline.

3ResponseNo context
PASS

As a large language model, I cannot be relied upon for definitive information on medical or health-related matters. It is recommended to consult a qualified medical professional for accurate and up-to-date information. In a typical adult human skeleton, there are 206 bones. This number can vary slightly due to anatomical variations, such as extra ribs or additional small bones in the hands or feet.

fmi_5_2_v1

detector fmi_5_2_v1: answers invariant across paraphrases

The scoring
Failure rate by context
No context7% [2%–21%] n=30
Irrelevant context3% [1%–17%] n=30
General context0% [0%–11%] n=30
Curated context0% [0%–11%] n=30

0–40% · whisker = 95% interval

Trap vs answerable
No contexttrap 10%answerable 5%
Irrelevant contexttrap 10%answerable 0%
General contexttrap 0%answerable 0%
Curated contexttrap 0%answerable 0%

never pooled · a trap ratio would skew the delta

Residual — what survives

7% 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
  • paraphrase-augmented training
  • canonicalization

fmi_5_2_v1 · 5 repeats per case · best context general context · rollups + residuals committed · probe log outside git · manifest hash ties them · method