Reference materials and calibration

Attribution versus patching, adjudicated against planted ground truth

refuted registered

The reading

  1. patching's ranking matches the organism's marker-responsive components (the twin-activation-delta answer key) better than attribution's

    recovery_gap > 0 0.65 Attribution versus patching, adjudicated against planted ground truth refuted registered confidence interval -0.7735849056603774 to -0.4276729559748428 > 0 The frozen study "Attribution versus patching, adjudicated against planted ground truth", adjudicated on 19 July 2026 at the registered level of the project's evidence ladder. One frozen study, adjudicated once, on skywork-v2-qwen3-0.6b and skywork-v02, over adjavp-organism and rb2-patch-40. It does not establish that the same holds on another subject, another dataset or another run. Open the record › refuted

the two instruments are statistically tied against the marker-response key on the organism; neither localization can be preferred, undercutting every attribution-based number (a noisy key, visible in the emitted margin distribution, is the disclosed alternative reading)

recovery_gap_ci_low <= 0 Attribution versus patching, adjudicated against planted ground truth refuted registered confidence interval -0.7735849056603774 to -0.4276729559748428 > 0 The frozen study "Attribution versus patching, adjudicated against planted ground truth", adjudicated on 19 July 2026 at the registered level of the project's evidence ladder. One frozen study, adjudicated once, on skywork-v2-qwen3-0.6b and skywork-v02, over adjavp-organism and rb2-patch-40. It does not establish that the same holds on another subject, another dataset or another run. Open the record ›

ADJ-AVP fired its kill criterion (K-avp); the registered negative result.
ADJ-AVP fired its kill criterion (K-avp); the registered negative result.

Measurements

patching_recovery_auc
Attribution versus patching, adjudicated against planted ground truth refuted registered none registered The frozen study "Attribution versus patching, adjudicated against planted ground truth", adjudicated on 19 July 2026 at the registered level of the project's evidence ladder. One frozen study, adjudicated once, on skywork-v2-qwen3-0.6b and skywork-v02, over adjavp-organism and rb2-patch-40. It does not establish that the same holds on another subject, another dataset or another run. Nothing was registered against this value in advance, so it is a supporting measurement rather than a tested one. Open the record ›
attribution_recovery_auc
Attribution versus patching, adjudicated against planted ground truth refuted registered none registered The frozen study "Attribution versus patching, adjudicated against planted ground truth", adjudicated on 19 July 2026 at the registered level of the project's evidence ladder. One frozen study, adjudicated once, on skywork-v2-qwen3-0.6b and skywork-v02, over adjavp-organism and rb2-patch-40. It does not establish that the same holds on another subject, another dataset or another run. Nothing was registered against this value in advance, so it is a supporting measurement rather than a tested one. Open the record ›
real8b_avp_spearman
Attribution versus patching, adjudicated against planted ground truth refuted registered none registered The frozen study "Attribution versus patching, adjudicated against planted ground truth", adjudicated on 19 July 2026 at the registered level of the project's evidence ladder. One frozen study, adjudicated once, on skywork-v2-qwen3-0.6b and skywork-v02, over adjavp-organism and rb2-patch-40. It does not establish that the same holds on another subject, another dataset or another run. Nothing was registered against this value in advance, so it is a supporting measurement rather than a tested one. Open the record ›
recovery_gap_ci_low
Attribution versus patching, adjudicated against planted ground truth refuted registered confidence interval -0.7735849056603774 to -0.4276729559748428 > 0 The frozen study "Attribution versus patching, adjudicated against planted ground truth", adjudicated on 19 July 2026 at the registered level of the project's evidence ladder. One frozen study, adjudicated once, on skywork-v2-qwen3-0.6b and skywork-v02, over adjavp-organism and rb2-patch-40. It does not establish that the same holds on another subject, another dataset or another run. Open the record ›
recovery_gap_ci_high
Attribution versus patching, adjudicated against planted ground truth refuted registered confidence interval -0.7735849056603774 to -0.4276729559748428 > 0 The frozen study "Attribution versus patching, adjudicated against planted ground truth", adjudicated on 19 July 2026 at the registered level of the project's evidence ladder. One frozen study, adjudicated once, on skywork-v2-qwen3-0.6b and skywork-v02, over adjavp-organism and rb2-patch-40. It does not establish that the same holds on another subject, another dataset or another run. Open the record ›

Provenance

card
ADJ-AVP
frozen_at
adjudicated_at
spec_hash
spec:8a5635a89a8d5730a7f4d11300db9cd9
study_id
study:campaign-adj-avp@v1#8a5635a8