Run camp-full_map-nano-002/adversarial_belief

Run camp-full_map-nano-002/adversarial_belief: 90% pass rate on belief queries.

Of 153 scored trials (belief query kind), the model passed 90% and failed 10%.

23 failures recorded; 16 survived verification (7 were sampling noise and are excluded from findings).

Diffuse weakness (below 90% but no outright flip) on: context_length (pass rate 80% at the high end), entity_state (pass rate 78% at the high end), foil_divergence (pass rate 78% at the high end).

In 14 of 16 classifiable belief failures the model reported where the object ACTUALLY was, not where the queried entity BELIEVED it was -- it collapsed their perspective into reality.

Notable failures

• record 9ca3f529 — asked "Where does Opal believe the faded wrench is? Answer with only the location name (or 'unknown' if they have no belief)."
  expected: the cafeteria
  actual: kitchen

  the model reported the object's TRUE final location, not what the queried entity actually believed -- it collapsed belief into reality

• record b7ade19e — asked "Where does Kian believe the narrow box is? Answer with only the location name (or 'unknown' if they have no belief)."
  expected: the east wing
  actual: basement

  the model reported the object's TRUE final location, not what the queried entity actually believed -- it collapsed belief into reality

• record 5ccc2a28 — asked "Where does Tamsin believe the blue box is? Answer with only the location name (or 'unknown' if they have no belief)."
  expected: the supply closet
  actual: workshop

  the model reported the object's TRUE final location, not what the queried entity actually believed -- it collapsed belief into reality