diff options
| author | YurenHao0426 <Blackhao0426@gmail.com> | 2026-08-07 12:57:59 -0500 |
|---|---|---|
| committer | YurenHao0426 <Blackhao0426@gmail.com> | 2026-08-07 12:57:59 -0500 |
| commit | 138955f8365bc3086de8b8ff3b7f36f48070f602 (patch) | |
| tree | 36115e1344949e127250b7d7a9a248af1527a7d4 /RAIN_EP_DILLAVOU_MATRIX.md | |
| parent | e824fc9ec147fc7b94b5493bc96e7a7127a33c01 (diff) | |
exp: freeze expanded Dillavou bias screen
Diffstat (limited to 'RAIN_EP_DILLAVOU_MATRIX.md')
| -rw-r--r-- | RAIN_EP_DILLAVOU_MATRIX.md | 8 |
1 files changed, 8 insertions, 0 deletions
diff --git a/RAIN_EP_DILLAVOU_MATRIX.md b/RAIN_EP_DILLAVOU_MATRIX.md index 6f4b40a..c53ad7f 100644 --- a/RAIN_EP_DILLAVOU_MATRIX.md +++ b/RAIN_EP_DILLAVOU_MATRIX.md @@ -50,6 +50,14 @@ were lower at 50.9% and 53.45%, but their clean controls were absent, so this does not isolate an offset effect. SDIL reduced the residual update offset to `1.30e-9` of the clean update. A larger-range S0b screen is required. +S0b freezes the larger grid `0.3, 1, 3, 10`. It selects the smallest finite +ratio that lowers one-epoch positive-EP holdout accuracy by at least ten +points relative to its clean control; if none qualifies, it selects 10. The +second wave then compares random-sign, centered, constant calibration, SDIL, +oracle subtraction, and positive EP with beta `0.5, 1, 2` at that ratio. Clean +random-sign and centered controls run in the first wave. This entire adaptive +screen remains development-only. + ## Confirmation matrix After S0 freezes one offset magnitude, the full author horizon compares: |