← Complete research archive
Language compilerClosed / no-go67 lines

R12 Structured Complete-Compiler Diagnostic Result

Protocol: r12 referential literal pointer compiler v1 2 structured development Decision: REJECT AS AN INTEGRATED COMPILER; RETAIN STRUCTURAL EFFECT

R12_REFERENTIAL_LITERAL_POINTER_COMPILER_STRUCTURED_DIAGNOSTIC_RESULT.mdOpen original Markdown ↗

R12 Structured Complete-Compiler Diagnostic Result

Protocol: r12_referential_literal_pointer_compiler_v1_2_structured_development Decision: REJECT AS AN INTEGRATED COMPILER; RETAIN STRUCTURAL EFFECT

Run identity

Job 692983 completed on a verified H100 PCIe on evc25 with exit code zero in 13m28s. The fit used the frozen v1.1 train/development bytes, one epoch, 1,514 updates, seed 2026071804, and the frozen optimizer schedule. Shohin remained frozen.

ItemValue
adapter parameters6,402,701
total parameters131,484,365
fit elapsed555.566 seconds
initial adapter-state SHA-25678add721586562cac4418fc539d8495737d772260354532a7f59f8b467ecfe15
final adapter-state SHA-256dd5d5b6a8c7d300c3fe4016098795a8caa9d671ed72a1a25a148e8f51de6be5c
adapter file SHA-2568d2278c369ead039a48bc39d8f9effab7198a5cecfabaf01e3724eec4ec3aa11
development result SHA-2560f237c052102955c26cc14340d16fe2139a020cac959ac27720f2faf22121d03
log SHA-256822c646fd55abaecbe80bb805bee8c6eb99dfcfeb309e15a23aa28e359985293

All three artifacts are hash-matched between Newton and the Mac. Confirmation access is zero.

Frozen result

Metricv1.1 free slotsv1.2 structuredChange
initial-state joint exact18.848%48.340%+29.492pp
full pointer exact2.197%0%-2.197pp
semantic-program exact15.283%0%-15.283pp
answer accuracy29.395%18.994%-10.401pp
operation-kind accuracy96.265%49.927%-46.338pp
canonical + paraphrase both exact0/5120/5120

The structural field improves ordered initial binding materially. On the paraphrase renderer, initial-state joint rises to 60.547%, operation-0 joint to 40.625%, and operation-1 joint to 38.672%. But the operation-kind loss remains at chance throughout training, so no semantic program is exact.

Mechanism diagnosis

The role head can minimize pointer and role losses without forcing the free operation slot to read the selected kind token. Structural location and semantic classification became disconnected parameter paths. This is a wiring failure, not evidence against bidirectional parsing.

A post-hoc diagnostic composes v1.2 pointer predictions with the independently trained v1.1 kind predictions, without fitting new weights. It reaches:

Hybrid metricResult
answers987/2,048 = 48.193%
semantic programs691/2,048 = 33.740%
full pointer exact224/2,048 = 10.938%
paraphrase answers270/512 = 52.734%
paraphrase semantic programs224/512 = 43.750%
all-four answers correct114/512

This hybrid is an external diagnostic, not an autonomous compiler. It supplies evidence that the structural and semantic gains coexist when their parameter paths are isolated.

Consequence

Reject v1.2 as an integrated compiler. A minimal successor may give structural roles and operation semantics independent memory projections/decoders while retaining the same source-only inference boundary. Passing exposed development can authorize only a fresh board and matched controls; v1.1 confirmation remains sealed.