item 27 scored: 34.0 best-in-series with zero pauses, but paired ns (burst 10-6 p=0.45; vs positional 44-38 p=0.58) — ceiling holds; pause tape confirmed ~worthless (ablation 32.4)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-17 02:12:54 +02:00
parent 6542f80339
commit 67f9c3091c
4 changed files with 3131 additions and 0 deletions
@@ -780,6 +780,22 @@ after it (job renamed zzz_t).
    burst also encodes the step trajectory; encoding is never the
    obstacle. Accuracy pending.
 --- Outcome, item 27 (scored 2026-07-17 ~04:35). k=0 sanity clean.
 Matched cell 2:0 (burst on) = 34.0 (drop .19 / easy .621 / hard .394)
 — nominally the BEST d=1 number of the series (vs 31.6 positional,
 31.2 lens-pauses) with ZERO pause tokens; ablation (same adapter,
 burst off) = 32.4. But both pre-registered comparisons are
 noise-compatible when paired: burst-on vs burst-off 10-6 discordants,
 p=0.45; vs positional d=1 44-38, p=0.58. Verdict: within the +-5
 band — the ceiling holds on the internal-iteration axis too; the
 burst's causal contribution is directionally positive (all of it in
 the hard bucket, +3.2) but unproven at n=256. Notable anyway: the
 zero-pause FORMAT loses nothing (ablation 32.4 >= positional 31.6 with
 12 fewer latent positions) — the pause tape was contributing ~nothing,
 consistent with item 25's display-without-routing reading. Item 28
 (teacher-state targets, same architecture) runs next and completes
 the 2x2 (verbal/full-state x tape/no-tape).
 28. **Teacher-state distillation into the burst (pre-registered
    2026-07-17 ~02:30, before running; Nils's variant: "meaningful CoT
    chunks yield internal state that we then teacher-force into the