From 75bc8efd85b365f3ff8006bc738d9e8cb8a85f96 Mon Sep 17 00:00:00 2001 From: Ronald Tse Date: Sun, 6 Sep 2026 02:39:35 +0200 Subject: [PATCH] revert: stray RESULTS.md from a wrong-repo merge (#156) The GKD verdict entry belongs in interscript-ml's results ledger, not the website repo. --- docs/RESULTS.md | 20 -------------------- 1 file changed, 20 deletions(-) delete mode 100644 docs/RESULTS.md diff --git a/docs/RESULTS.md b/docs/RESULTS.md deleted file mode 100644 index 8b1fd38..0000000 --- a/docs/RESULTS.md +++ /dev/null @@ -1,20 +0,0 @@ - -## ara-diac-small-2-gkd — on-policy distillation NEGATIVE: 6.0036 (2026-09-06) - -The on-policy rung (GKD: student-generated mistakes scored by the -teacher, a120384 wiring, 10,995 steps, labels sha e70ce991): teacher -reproduces 2.289, student **6.0036** full-set, paired bootstrap delta -3.4083 [3.109, 3.743], p=0 — gate FAILED, and 1.43pp WORSE than the -off-policy sequence-KD rung it was meant to improve (G2a 4.5701). - -This completes the residual-attribution program with a clean pattern: -every lever tested fails to close the teacher-student gap — -* classical corpus add (G2b): flat-negative -* register swap (E6): negative -* on-policy distillation (GKD): negative, worse than off-policy -* memory layers (PKM): real but small (0.70pp), below bar -* epochs: small (0.25pp) -The residual is not data, not domain coverage, not an off/on-policy -deficit. It is a property of the compression itself at this rung — -the honest open question for the paper. On-policy stays listed in -Paper B's future work with its measured negative attached.