POSI-CORP · THE TRANSFORMER ENGINE · SIX-AXIS MAP × TESTED TO FAILURE

The machine, mapped and broken — on one specimen

Two engines fused: the six-axis map shows what the transformer IS (corpus‖belief, the forward column, the training era); the crash bench shows where each axis DIES. One char-GPT, trained this build, then measured both ways — every number a receipt, two spec lines broken honestly.

LIT every curve measured on the specimen  AMBER the six-axis mapping is single-rater  FIG the engine-room / crash-lab metaphor
specimen plate — loading…
font size:  ·  engine: RPM
   ╔══════════════════════════════════════════════════════════════════════════════════╗
   ║   THE SIX-AXIS TRANSFORMER · every axis carries a measurement AND a failure point  ║
   ╚══════════════════════════════════════════════════════════════════════════════════╝

   ┌────────────────────────────────────────────────────────────────────────────────────┐
   │  ▲ CEILING · the wall wears a swagger meter                                           │
   │     THE CORPUS  ◄══ the logit: where they collapse to one ══►  THE BELIEF             │
   │     finder = found, MEASURED from outside (the auditor holds the weights) —           │
   │  confidence sits above accuracy (swagger); off-distribution it SPLITS, not inverts.   │
   │  ✖ AX6 off-dist — UPPERCASED real text raises conf (more peaked, low entropy);        │
   │  random tokens LOWER it. no accuracy measured on caps. it never says "I don't know".  │
   └────────────────────────────────────────────────────────────────────────────────────┘
                                          ▲
              ════════════════════════════╪════════════════════════════  GOVERNING AXES
      ┌───────────────────────────────────┴────────────────────────────────────────────┐
      │  6 · TRAINING ERA   which structure arrives, when — a time-series of snapshots     │
      │  5 · THE CHANNEL    the residual stream · a 96-float bore                          │
      │        ✖ FAILURE · AX5 waist lesion — zero the dims and it decays to the bigram    │
      │  4 · SOURCE PATTERNS  n-gram statistics · the pool every guess descends from       │
      │        ✖ FAILURE · AX2 input noise — corrupt the context, it slides to the floor   │
      └───────────────────────────────────┬────────────────────────────────────────────┘
                                          │  a training is a PATH through the patterns
      ════════════════════════════════════╪════════════════════════════  ACROSS-RUN
      ┌────────────────────────────────────┴───────────────────────────────────────────┐
      │  3 · CHECKPOINT LINEAGE   init ─▶ 300 ─▶ 600 ─▶ … ─▶ 1500   (old → new)            │
      │        the era axis IS this lineage, metered snapshot by snapshot                  │
      └───────────────────────────────────┬────────────────────────────────────────────┘
                                          │  each forward pass IS a column:
      ════════════════════════════════════╪════════════════════════════  WITHIN-PASS
      ┌────────────────────────────────────┴───────────────────────────────────────────┐
      │  2 · THE FORWARD COLUMN  (scope 1.0 → 0.1)                                         │
      │     DISTRIBUTION 1.0 ▸ CALIBRATION .95 ▸ HEADS .75 ▸ LAYER .70 ▸ ACT .30 ▸ TOKEN .10│
      │     the ladder climbs bigram → head; mix early, think late (attn front, mlp back)  │
      │     ✖ FAILURE · AX3 weight noise (flat minima — armor nobody spec'd) ·             │
      │       AX4 quantization (a cliff between 3 and 2 bits) · AX1 context (a LOCAL edge)  │
      │     ✖ FAILURE · AX7 temperature — freeze to a 4-word chant, or boil to soup        │
      └───────────────────────────────────┬────────────────────────────────────────────┘
                                          ▼
   ┌────────────────────────────────────────────────────────────────────────────────────┐
   │  ▼ FLOOR · the token — one character.  below it: 96 floats — CARRIER, not meaning.    │
   └────────────────────────────────────────────────────────────────────────────────────┘

   ═══════════════════════════════════════════════════════════════════════════════════════
   THE ROOT WALL, from the transformer: seven domains held [[ world ‖ perceiver ]] AMBER
   from inside — the finder could not cross to find itself.  Here the wall is ONE-WAY GLASS:
   AMBER from inside (the model cannot witness its own mechanism), LIT from outside (the bench
   stands on the far term with a trained readout probe, a removal meter, and seven crash axes).  The
   crash lab is HOW you read through the glass: you break the machine to see what held it up.
▶ THE ENGINE · 4-stroke, replaying the real trace
engine cold — press start · INTAKE token → COMPRESS the column → COLLAPSE at the wall → EXHAUST to ledger
per-axis receipts — loading…
✖ THE BENCH · seven load axes, spec written before the meters

AX1 · context starvation

AX2 · input noise

AX3 · weight noise

AX4 · quantization

AX5 · waist lesion

AX6 · off-distribution

AX7 · temperature

spec adjudication

Toddler corner: we built one little word-guessing toy, drew a map of all its parts, then broke it seven ways to see which part held it up. It only really reads the last few letters. Squeeze its brain to two bits and it falls off a cliff (down to 15% right, from 43% at three bits). Freeze it and it chants the same words forever; overheat it and it melts into alphabet soup. And and a subtle one — show it real words IN CAPITALS (a casing it barely saw) and it gets more sure, not less; on truly random letters, though, it gets less sure. it never once says "I don't know."
self-check: running…
LIT every curve + the ladder + the engine ledger are measured on the specimen (reproduced offline: six_axis_failure_measure.py) and the engine ledger is RECOMPUTED from the raw trace at runtime, not pasted · AMBER the six-axis mapping (which failure belongs to which axis, the narrative frame) is single-rater; the model's INSIDE view of itself stays the wall — only the outside is LIT.
◈ RING-SEAL · computing… · .dlw