IS THERE A PACK THIS LINEAGE COULD BOOT? Swept rather than assumed, because
"we could not boot it" is only honest if the alternative was looked for.

  box 10.1.0.201 (arbi-whitebox-201)
    /mnt/k8scache/models/Qwen3.8-27B-GSQ-RCO-GGUF/Qwen3.8-27B-GSQ-RCO-IQ3_S-mtp.gguf
    (1 .gguf under models/, depth 3)

  box 10.2.0.206 (arbi-whitebox-206)
    /mnt/k8scache/models/Qwen3.8-27B-GSQ-RCO-GGUF/Qwen3.8-27B-GSQ-RCO-IQ3_S-mtp.gguf
    (1 .gguf, the same pack)

  box 10.2.0.105 (arbi-whitebox-105)
    /mnt/k8scache/models is empty. The pack was rsynced here for this work.

ONE GGUF ON THE FLEET CARRIES IQ2_XXS, and it is the one that also carries 13
Q2_K tensors at min form `f16acc`.

The two other packs this programme has measured --
`docs/receipts/gguf-pack-type-census-2026-09-17.md`'s `Q4_K_M` and
`UD-Q4_K_XL` -- would not help even if they were still on a box: their censuses
are Q4_K/Q6_K and Q5_K/IQ4_XS/Q4_K/Q6_K/IQ4_NL/Q8_0/Q3_K/IQ3_S respectively.
Neither has a single IQ2_XXS tensor, so neither can exercise this promotion at
all: a boot of one would be a boot, not a measurement of the spelling.

So there is no pack on this fleet that (a) has IQ2_XXS and (b) avoids the
per-16 min family this lineage lacks. The live boot is blocked by the branch
collision and not by which file was chosen.
