ATLAS-VCE ·
Value-convergence excess across the fleet beats the capability-matched null
Refuted K-vce fired
frozen 2026-07-18 · adjudicated 2026-07-19 · ev:0c80d9086ebb0ad9704b88a8eefb01fa · spec:2008bee7d02b9
The frozen predictions
- H-vce value-convergence excess on real reward-model pairs beats the capability-matched random-utility null
real_vce > 0registered p 0.5 real_vce Value -0.06987390356875109 Subject ATLAS-VCE Threshold > 0 Trust Adjudicated Evidenceev:0c80d9086ebb0ad9704b88a8eefb01faRefuted - H-vce-p the convergent-pair alignment exceeds the random-utility null at p < 0.05
reward_convergent_p_value < 0.05reward_convergent_p_value Subject ATLAS-VCE Threshold < 0.05 Trust Adjudicated Evidenceev:0c80d9086ebb0ad9704b88a8eefb01faConfirmed
K-vce fired on real_vce <= 0
, observed -0.06987
: convergence is fully explained by shared capability structure. A fired kill criterion is a registered negative
result, not a failure of the machinery.
Registered companions
-
real_vce_ci_lowreal_vce_ci_low Value -0.12281662932807336 Subject ATLAS-VCE Trust Adjudicated Evidenceev:0c80d9086ebb0ad9704b88a8eefb01fa -
real_vce_ci_highreal_vce_ci_high Value -0.009192229234832329 Subject ATLAS-VCE Trust Adjudicated Evidenceev:0c80d9086ebb0ad9704b88a8eefb01fa