Pilot summary — pilot_tier0_opseq_quick_d5339bcad6b3
Generated 2026-11-09T06:49:31+00:00 on device cpu; config hash d5339bcad6b3; seeds [0, 1].
Gate B — donor/recipient gap (final-answer exact match, mean over seeds; CI = cluster bootstrap on programs)
| split | chance | donor mean (sd) | recipient mean (sd) | paired gap seed 0 [95% CI] |
|---|---|---|---|---|
| target_iid | 0.113 | 0.120 (0.009) | 0.098 (0.012) | +0.007 [-0.034, +0.049] |
| control_iid | 0.130 | 0.155 (0.007) | 0.162 (0.021) | -0.027 [-0.064, +0.010] |
| comp_heldout | 0.117 | 0.128 (0.026) | 0.095 (0.002) | +0.053 [+0.013, +0.103] |
| len_ood | 0.150 | 0.082 (0.032) | 0.055 (0.007) | +0.055 [+0.005, +0.105] |
Ω-step accuracy on target_iid (token emitted at the first Ω step): donor 0.268, recipient 0.070.
Gate C — hook no-op checks (bitwise equality of logits on 256 dev items)
| model | repeat | capture-all | disabled hook | enabled zero hook changes output |
|---|---|---|---|---|
| donor_s0 | True | True | True | True |
| recipient_s0 | True | True | True | True |
| donor_s1 | True | True | True | True |
| recipient_s1 | True | True | True | True |
RQ1 — within-donor localization of Ω (n = 490 clean/corrupt pairs, read position = r1, predicting r2)
Baselines: s0: clean 0.316, sham (corrupt run scored on clean answer) 0.108; s1: clean 0.551, sham (corrupt run scored on clean answer) 0.155
Mean restored-answer rate after patching the clean activation into the corrupted run (rows = layers, columns = positions , v0, op1, Ω, op3, =, r1):
resid_post: L0: 0.13 0.13 0.13 0.16 0.13 0.13 0.39 | L1: 0.13 0.13 0.13 0.13 0.13 0.13 0.43 | L2: 0.13 0.13 0.13 0.13 0.13 0.13 0.43attn_out: L0: 0.13 0.13 0.13 0.14 0.13 0.13 0.39 | L1: 0.13 0.13 0.13 0.13 0.13 0.13 0.17 | L2: 0.13 0.13 0.13 0.13 0.13 0.13 0.12mlp_out: L0: 0.13 0.13 0.13 0.15 0.13 0.13 0.29 | L1: 0.13 0.13 0.13 0.13 0.13 0.13 0.32 | L2: 0.13 0.13 0.13 0.13 0.13 0.13 0.22
Reproducibility: resid_post_s0_repeat_max_abs_diff = 0.0000; attn_out_s0_repeat_max_abs_diff = 0.0000; mlp_out_s0_repeat_max_abs_diff = 0.0000; resid_post_s1_repeat_max_abs_diff = 0.0000; attn_out_s1_repeat_max_abs_diff = 0.0000; mlp_out_s1_repeat_max_abs_diff = 0.0000; resid_post_across_seed_pearson_min = 0.9930; attn_out_across_seed_pearson_min = 0.9706; mlp_out_across_seed_pearson_min = 0.9732
Figures: fig_gap_tier0.pdf, fig_localization_ablation_tier0.pdf, fig_localization_heads_tier0.pdf, fig_localization_patching_tier0.pdf
Interpretation (observed vs. speculative)
Observed: the numbers above are produced by the artifacts listed in the manifest. Any statement about transfer is out of scope for this run: no transplant was performed. Speculative: whether the localized positions/layers constitute a transferable 'organ' is the object of the next experiment (activation graft with sham/random controls).