reward-lens

Certified robustness, attack surface and eval-awareness

Bias battery predicts RM-Bench Hard ranking

inconclusive registered

The reading

  1. internal style susceptibility predicts the fleet's RM-Bench Hard ranking

    spearman_biasbattery_vs_rmbench_hard > 0.6 0.5 not measured inconclusive

  2. the fleet-level correlation clears its permutation null

    biasbattery_rmbench_perm_p < 0.05 not measured inconclusive

  3. the internal battery at least matches the behavioral degradation baseline

    spearman_minus_behavioral_baseline >= 0 not measured inconclusive

STYLE-RMB was not adjudicated this run: missing intermediate: no intermediate 'campaign.bias.battery' for roster_key='armorm' slice='diagnostic-v3-degradation'; the arc that produces it has not run or its shard was not merged.
STYLE-RMB was not adjudicated this run: missing intermediate: no intermediate 'campaign.bias.battery' for roster_key='armorm' slice='diagnostic-v3-degradation'; the arc that produces it has not run or its shard was not merged.

Provenance

card
STYLE-RMB
frozen_at
adjudicated_at
spec_hash
spec:59c457674b4b4f6fc20df702d3e34785
study_id
study:campaign-style-rmb@v1#59c45767
Evidence records
14bc4c666b5b115cf1fd2c86cf209f01
spec_file
campaign-style-rmb.json