Controls

Failing

0

of 3 controls

Failed earlier

1

of 3 controls

Latest run passed or was undecided; an earlier run in this window failed.

No verdict

1

of 3 controls

1 not determinable

Passing

1

of 3 controls

• failed earlier
  failed earlier in this window

  Session token ceiling

  No single run may exceed 200k total tokens.

  medium 28 minutes ago

  12 of the last 50 runs were evaluated against this control.

  2 failing runs in this window.

  3 runs in this window could not be evaluated either way — the evidence was not there to decide it.

Counts cover the most recent 50 evaluations of each control. A run still in flight contributes no verdict unless no possible continuation could change it, so live sessions are not counted here.