datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
killinchu-osint-corpus
killinchu — live intel archive
Append-only, content-addressed archive of the live intelligence streams the
killinchu demo ingests, published by SZL Holdings. The shared
szl_hf_bucket client writes one NDJSON shard per UTC day under
intel/.
Each row is {schema, id, ts, source, kind, payload}.
Viewer-safe contract
The default Dataset Viewer configuration is the homogeneous
archive_manifest, one row per immutable raw shard. It exposes path, byte
count, Git blob hash… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/killinchu-osint-corpus.szl-lake
🪢 SZL Holdings Data Lake (szl-lake)
A public, signed, diligence-defensible corpus of governance provenance.
Every governed action in the SZL substrate emits an ECDSA P-256 DSSE-signed Khipu receipt onto a hash-linked Merkle DAG. This dataset aggregates those receipts together with formal-verification doctrine snapshots, the Zenodo paper record, SBOM pointers, and compliance attestations so a reviewer can verify the claims end-to-end.
Doctrine v11 LOCKED · 749… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-lake.a11oy-verifiable-corpus
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
a11oy — Verifiable Corpus · verify it yourself
This dataset publishes a11oy's signed receipts and proof surface so that
anyone can independently verify them — no trust in SZL Holdings required.
Every receipt here carries the full cryptographic material needed to check its
signature offline;… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/a11oy-verifiable-corpus.szl-artifacts
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
SZL Artifacts — Build Artifact Registry
Artifact boundary - audited 2026-07-15: this Hugging Face dataset repository
is a mixed 106-file, 36,100,909-byte publication/build mirror at revision
91bcb443857f2884ef2bfabaaa6bfdc606c7134a. It is not a uniform table of
DSSE envelopes, and… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-artifacts.uds-governance-receipts
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
UDS Governance Receipts — Decision Audit Log
Doctrine v11 LOCKED. No marketing. Every number resolves to a CI log, a Lean proof, or a Zenodo DOI.
Append-only log of DSSE-signed governance decision receipts for the Unified Deployment Substrate (UDS) mesh. Each record captures:… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/uds-governance-receipts.lean-proofs-v1
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
SZLHOLDINGS/lean-proofs-v1
The complete Lean 4 theorem library for the SZL Holdings Ouroboros Invariant research programme.
Doctrine v10/v11 Canonical Numbers
Metric
Value
Declarations
749
Unique axioms
14 (15 raw, 1 dup)
Sorries
163 (112 baseline + 51 Putnam)… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/lean-proofs-v1.uds-spans-receipts
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
UDS Spans Receipts — OTel Governance Audit Log
Doctrine v11 LOCKED. No marketing. Every number resolves to a CI log, a Lean proof, or a Zenodo DOI.
Append-only audit log of DSSE-signed OpenTelemetry spans emitted by the UDS mesh governance layer. Each span record includes: operation… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/uds-spans-receipts.thesis-v18-formal-verification
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
Ouroboros Thesis v18 — Formal Verification
Doctrine v11 LOCKED. No marketing. Every number resolves to a CI log, a Lean proof, or a Zenodo DOI.
Historical snapshot — this dataset is the v18-specific Lean mechanization index. The live source of truth is lean-proofs-v1, which is kept… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/thesis-v18-formal-verification.szl-frontier-evaluation-receipts
Checkable language-model evaluation records
We publish the inputs, measured responses, scoring results and file hashes from small language-model evaluations. These records help readers inspect what was tested and check whether files changed.
Archive and Viewer scope
The default Dataset Viewer configuration selects only runs/**/summary.json: one row per published evaluation run, not one row per test case. The train split is an archive indexing label, not a training… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-frontier-evaluation-receipts.thesis-corpus-v18
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
SZLHOLDINGS/thesis-corpus-v18
The v18 Ouroboros Invariant thesis — LaTeX chapters, the 179 formal blocks
(theorem / lemma / definition / axiom environments) as a flat CSV, and the per-version
delta ledger that tracks how every formal block evolved v1 → v18.
Contents
File… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/thesis-corpus-v18.doctrine-v10-v11
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
SZLHOLDINGS/doctrine-v10-v11
The locked governance doctrine for the SZL Holdings agentic substrate: Doctrine v10,
Doctrine v11 (13-axis yuyay_v3 canonical, supersedes v10), the Mythos → Hatun-Willay
rename rule, and the canonical wedge.
Contents
File
What… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/doctrine-v10-v11.why-we-lead
Why We Lead — HISTORICAL (Doctrine v7, not current)
HISTORICAL. This dataset is a Doctrine v7 positioning snapshot. It is not current Doctrine v11. Do not present any count in these files as current MEASURED. Images and text are frozen audit copies.
An investor opening this ID should treat every metric as a SNAPSHOT of an old document, not the live org.
What it is / is NOT
IS: a historical positioning corpus (Doctrine v7 era) plus images under images/.
NOT:… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/why-we-lead.canonical-formulas-v1
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
SZLHOLDINGS/canonical-formulas-v1
The canonical SZL formula registry — 21 pure, typed, no-IO Python formulas, the
matching Lean 4 obligation theorems, and the Codex-Kernel governed-loop composer.
Contents
File
What
code/python/formulas.py
21 canonical formulas, each… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/canonical-formulas-v1.readiness-runs
SZL Readiness Runs
Signed Khipu receipts emitted by the eight SZL Production-Readiness agents.
Doctrine v11 (LOCKED): 749 / 14 / 163.
Layout
receipts/<agent>/<UTC-date>/<UTC-timestamp>.json # one DSSE envelope per run
dr-dumps/<flagship>/<ts>.ndjson # READINESS-DR backup dumps
Agents: readiness-reliability, readiness-security, readiness-observability,
readiness-operability, readiness-compliance, readiness-docs, readiness-dr… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/readiness-runs.uds-bundles-v1
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
SZL UDS Bundles v1
Airgap-deployable UDS Core / Zarf productionization artifacts for the SZL Holdings organ substrate (a11oy, amaru, sentra, killinchu, rosie).
Contents
uds_productionization/PER_BUNDLE/<organ>/ — per-organ Dockerfile, Helm chart, and UDS manifests (namespace… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/uds-bundles-v1.model-bom
SZLHOLDINGS Model BOM Registry
Canonical CycloneDX 1.5 Model BOM registry for all 44 public models under the
SZLHOLDINGS organization.
Provenance
Generated 2026-08-30 by szl_estate_audit from live Hugging Face Hub metadata
(model listings, repository file siblings, and card data). Each BOM is a
pretty-printed CycloneDX 1.5 JSON document carrying a unique urn:uuid
serial number.
File layout
bom-index.json - machine-readable index: one entry per… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/model-bom.szl-frontier-covenant
SZL Frontier — Memory Covenant v0.3
Software release artifacts for the SZL Frontier orchestrator: policy formulas, software gates, honest posture labels, and governed frontier-evaluation intake records.
This is not a training corpus of classified or operational intelligence. Intel observations in the companion app are simulated and public-style only. The frontier intake file stores metadata, source revisions/fingerprints, evaluation lanes, and promotion posture; it does not… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-frontier-covenant.rag-corpus-v1
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
SZLHOLDINGS/rag-corpus-v1 — Agentic-RAG corpus + per-organ FAISS indexes
Doctrine v10/v11. Embedding model: BAAI/bge-base-en-v1.5 (768-dim).
Built by the agentic-RAG SHIP directive (390_AGENTIC_RAG_FAISS_PER_SPACE).
Contents
corpus.jsonl — 762 chunks, each ~512 tokens with… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/rag-corpus-v1.ouroboros-arxiv-preprint
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
Ouroboros arXiv Preprint — Thesis Draft
Doctrine v11 LOCKED. No marketing. Every number resolves to a CI log, a Lean proof, or a Zenodo DOI.
arXiv preprint artifact for the Ouroboros agentic-AI governance thesis. Canonical DOI: 10.5281/zenodo.20434276. No arXiv submission ID exists… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/ouroboros-arxiv-preprint.test-results
SZL test-results — anatomy alive-harness sink
Public, DSSE-signed results sink for the SZL anatomy alive-harness. This
dataset was retired earlier in 2026 and stood back up on 2026-07-21 as the
harness's fail-closed publishing target — restoring the public proof loop
behind every "harness verified" claim in the estate.
What a run is
anatomy_alive_v6.py (in this repo) drives live assertions across the whole
substrate — organ liveness, live formula-gate executions… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/test-results.governed-agent-bench
governed-agent-bench v0
This dataset is the immutable public mirror of
szl-holdings/a11oy@1b40fcbe0f65c1ad1e07776f83aa01abb067e864.
It measures five governability axes:
fail-closed behavior;
non-increasing authority across delegation;
false-success rejection;
receipt completeness; and
rollback discipline.
Evidence labels
Corpus: SAMPLE
Scores: COMPUTED
Receipt verification: STRUCTURE_ONLY
Cryptographic verification: false
The reference result proves that the… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/governed-agent-bench.szl-1-doctrine-sft
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.
SZL-1 Doctrine SFT
The supervised fine-tuning (SFT) set that teaches SZL-1 its identity and
SZL Holdings' honesty doctrine. This is the exact training data used by the
szl-forge kit (train_szl.py,
Unsloth QLoRA on unsloth/Qwen2.5-3B-Instruct-bnb-4bit).
What's in it — honest labels… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-1-doctrine-sft.szl-quant-sft-v1
szl-quant-sft-v1 — training rows with signed lineage
Every row in this dataset is derived deterministically from a DSSE-signed backtest receipt and is recomputable bit-exact from content-addressed archives. No row was hand-written, scraped, or synthesized by a model.
Lineage (verifiable end-to-end)
CoinGecko daily closes (REPORTED venue feed)
→ szl-quant MEASURED walk-forward backtests → DSSE-signed receipts… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-quant-sft-v1.receipted-unsloth
Receipted Unsloth
How SZL Holdings actually trains. Silhouette from Unsloth QLoRA. Cut is original SZL. We do not republish Unsloth Studio, Desktop, copy, code, or someone else's tensors.
Collection: Receipted Unsloth — LIVE
The house loop
Disclose the Apache base (Qwen/Qwen2.5-* or Qwen/Qwen3.5-0.8B).
Train with Unsloth FastLanguageModel QLoRA on owner metal or HF Jobs (uv run + HF_TOKEN).
Bind dataset SHA-256, LoRA knobs, seed, and loss into a training receipt.… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/receipted-unsloth.david-leads-data
David Leads — verified Federal Refresh
This dataset holds scheduled public-data inputs for the David Leads broker-research demo. The EPA lane uses the official weekly ECHO Exporter. DOL uses published Form 5500 disclosures; FMCSA and USAspending use bounded captures from their official APIs. Coverage is declared per snapshot.
Each lane is available only after verified publication: latest/echo-exporter.json, latest/fmcsa.json, latest/form5500.json, and latest/usaspending.json.… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/david-leads-data.SZLHOLDINGS
SZL Holdings - organization profile metadata
Lifecycle: INFORMATIONAL_METADATA_ONLY. Production-ready: false. At exact
upstream revision fdca3ff2d26bc25080a87a75c02994055f2e01bb, this repository
contains three metadata files and no dataset payload, records, schema,
configuration, or generated artifact. It must not be treated as training data.
Evidence and machine-readable status
provenance.json binds this candidate to the exact
Hugging Face dataset repository… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/SZLHOLDINGS.alloy-sovereign-eval-runs
Alloy Sovereign Eval Runs · the honest first measured run
Append-only measured eval runs produced by routing SZL's
K-Verify Benchmark v1
through the live Alloy governed-inference stack on SZL's own sovereign
metal (provider: sovereign, zero cloud, zero spend). Each row is one
inference: its verdict, latency, NVML-measured energy, and a
signed receipt id that is re-checkable against the live Alloy receipt chain.
Built and maintained by SZL Holdings. Apache-2.0.… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/alloy-sovereign-eval-runs.szl-estate-graph
SZL Estate Graph v1
This is a deterministic, local publication bundle for SZLHOLDINGS/szl-estate-graph.
It turns the two receipted SZL Constellation topology documents into a typed,
trinity-connected graph suitable for graph-learning and drift comparison.
Publication target: SZLHOLDINGS/szl-estate-graph. Hub publication and its
immutable revision are provider evidence separate from this content bundle;
neither publication nor download implies model training, admission, or… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-estate-graph.oac-clinical-transport-observability-synthetic
OAC Clinical Transport Observability — Synthetic
This dataset contains 1,200 fixed-seed, entirely synthetic operational
transport-health examples for the companion OAC System Health v1 model.
It contains no records collected from a patient, laboratory, analyzer,
instrument, LIS, EHR, network, or health-care site.
Companion model: OAC System Health v1.
Canonical source: szl-forge clinical gateway.
Data boundary
The closed schema contains only eight bounded… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/oac-clinical-transport-observability-synthetic.szl-second-brain-inrepo
SZL Second Brain — in-repo lane (public projection)
This is DATA, not a model. A retrieval corpus — the in-repo lane of the
SZL Second Brain — published as a Hugging Face dataset. It is part of the
SZL Forge Family. Publishing it is repo-existence of a corpus only: it
trains nothing, evaluates nothing, serves no inference, and upgrades nothing
(Λ stays Conjecture-1).
What this is
A deterministic BM25-ready corpus assembled only from repo-public text of
the Alloy… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-second-brain-inrepo.