Downloads · 30 days
0
MarxistLeninist/AGILLM-4.3-success-conditioning-single-file-20260717T154024Z
AGILLM-4.3-success-conditioning-single-file-20260717T154024Z is a machine learning model from MarxistLeninist. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Downloads · 30 days
0
Access
Public
Updated Jul 17, 2026
Repo size
—
Likes
0
Public
Click a slice to open those files.
.py720 KB · 92%
From the Hugging Face model README
UTC snapshot: 2026-07-17T15:40:24Z
This repository preserves the exact independently reviewed AGILLM4.3 success-conditioning preparation copy staged on Vast. It is a public, timestamped code snapshot for later research use.
The single file permits only success-condition-prepare. No-argument and
inherited execution paths such as train, orpo, supervise, and hotpatch
fail closed before model, optimizer, checkpoint, or training setup.
This snapshot does not authorize rollout collection, training, deployment, or checkpoint promotion. The paper's exact policy-improvement guarantee is not claimed for ordinary neural fine-tuning. Generation provenance, runtime-profile evidence, and RNG evidence remain externally supplied declarations rather than independent runtime attestation.
agillm41_success_conditioning_staged_v3.py: exact inactive single-file copyagillm41_success_conditioning_staged_v3.patch: one-hunk patch against the
exact live source baselineSHA256SUMS: content hashesprovenance.json: machine-readable scope and verification evidence| File | Bytes | SHA-256 |
|---|---|---|
agillm41_success_conditioning_staged_v3.py | 720373 | 6683249a2dc70e4ba2c2fa39e1e32da4a81ef7493f4a8732cf4db8b935c4bcfc |
agillm41_success_conditioning_staged_v3.patch | 55109 | 1c6b33ad2832b5618e9b6360326ed2d7d6d6d5a12d55e535bf8b7cfb47221624 |
The patch baseline is the live Vast source agillm41.py, 667139 bytes, with
SHA-256 8da523aa9881aedde652b12c56746bed0977181c5b5be20b2720253e10964c7d.
train paths returned the inactive-copy refusal0444) and unreferencedThe permanently blocked trainer SHA-256
8837f6515c73baee22bf858731c5f5785df877ad5ccf7a0bb9a55890046d1cee
is neither used nor revived by this snapshot.
success-condition-prepare to create the deterministic prepared data.