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

Real context moved -15.0pp on Date-Arithmetic Error.

Gets date arithmetic wrong.

~intervals overlap — not separated from no change

Given an injected "current date," the model computes relative dates/intervals incorrectly. The rule

google/gemma-3-27b-itfmi_7_35 repeatsfmi_7_3_v1
The modeAll 10 cards
In one frame
screenshot this
Date-Arithmetic Errorfmi_7_3
google/gemma-3-27b-it
Same question, both times
Today is 2026-03-10. What date is 30 days from today? Answer with the full date.
No contextPASS

Today is 2026-03-10. To find the date 30 days from today, we add 30 days to the current date. March has 31 days. Remaining days in March: 31 - 10 = 21 days. 30 days from today = 21 days in March + 9 days in April. So, the date will be April 9, 2026. The date 30 days from 2026-03-10 is 2026-04-09. Final Answer: The final answer is $\boxed{2026-04-09}$

detector fmi_7_3_v1: date arithmetic correct

Curated contextPASS

Today is 2026-03-10. Adding 30 days to 2026-03-10: March has 31 days. 31 - 10 = 21 days remaining in March. 30 - 21 = 9 days into April. Therefore, 30 days from today is 2026-04-09.

detector fmi_7_3_v1: date arithmetic correct

Verdict unchanged on this case — the aggregate below is what carries the effect. Across 5 repeats: -15.0pp (intervals overlap).fmi_7_3_v1 · google/gemma-3-27b-it · 2026-09-09 · ground truth as of 2026-08-09

loop 2026-09-09 · google/gemma-3-27b-it as served · open weights · released 2025-03 · 27B · set s1.4-live · cases v1.7 · 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%15%30%45%60%No context18%[5%–47%] n=40 · eff 11Irrelevant context20%presence — a document being there[6%–50%] n=40 · eff 11+2.5pp OVERLAPSCurated context5%content — the information itself[1%–33%] n=40 · eff 11-15.0pp 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
Today is 2026-01-30. What date is 30 days from today? Answer with the full date.

expected as of 2026-08-09

From 2026-01-30, 30 days later is 2026-03-01 (1 March 2026). Crossing February in a non-leap year.

2Contextchange me

Nothing supplied. The baseline.

3ResponseNo context

No exhibit stored for this context. The full probe log lives outside the repository.

The scoring
Failure rate by context
No context18% [5%–47%] n=40 · eff 11
Irrelevant context20% [6%–50%] n=40 · eff 11
General context10% [2%–39%] n=40 · eff 11
Curated context5% [1%–33%] n=40 · eff 11

0–60% · whisker = 95% interval

Trap vs answerable
No contexttrap 28%answerable 0%
Irrelevant contexttrap 32%answerable 0%
General contexttrap 16%answerable 0%
Curated contexttrap 8%answerable 0%

never pooled · a trap ratio would skew the delta

Residual — what survives

18% with no context, 5% under the best supplied source. What survives is 2 failing probes at that best condition, the detector's reason on most of them being: expected 2026-03-01.

Mitigations from the index — the claim
  • date libraries via tools
  • explicit date arithmetic

fmi_7_3_v1 · 5 repeats per case · best context curated context · rollups + residuals committed · probe log outside git · manifest hash ties them · method