Datasets:

SZLHOLDINGS
/

szl-quant-sft-v1

Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 290, in _generate_tables
                  pa_table = paj.read_json(
                      io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
                  )
                File "pyarrow/_json.pyx", line 342, in pyarrow._json.read_json
                File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
                  return check_status(status)
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
                  raise convert_status(status)
              pyarrow.lib.ArrowInvalid: JSON parse error: Column(/fields/deployment_evidence/evidence/[]/value) changed from string to boolean in row 0
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                                         ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 101, in _split_generators
                  pa_table = next(iter(self._generate_tables(**splits[0].gen_kwargs, allow_full_read=False)))[1]
                             ~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 304, in _generate_tables
                  batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
                  examples = [ujson_loads(line) for line in original_batch.splitlines()]
                              ~~~~~~~~~~~^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
                  return pd.io.json.ujson_loads(*args, **kwargs)
                         ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
              ValueError: Expected object or value
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                               ~~~~~~~~~~~~~~~~~~~~~~~^
                      path=dataset,
                      ^^^^^^^^^^^^^
                      config_name=config,
                      ^^^^^^^^^^^^^^^^^^^
                      token=hf_token,
                      ^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
                  info = get_dataset_config_info(
                      path,
                  ...<6 lines>...
                      **config_kwargs,
                  )
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

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
      github.com/szl-holdings/szl-quant/receipts/backtest_*.receipt.json
  → content-addressed input archives  data/datasets/<sha256>.json
  → tools/sft-export.mjs — deterministic replay of the engine's actual
    decisions (signature-verified receipts only; fail-closed on any mismatch)
  → quant_sft_v1.jsonl  (this repo)
  → quant_sft_v1.manifest.json + DSSE receipt over the manifest

Each row's provenance block cites: source receipt path, receipt file sha256, dataset archive sha256, bar index, generator. The signed manifest pins the JSONL bytes (jsonlSha256).

What the rows teach

Doctrine-governed advisory reasoning, in the engine's own voice:

honesty labels on every value (REPORTED feed, HEURISTIC components, advisory Λ);
conviction hard-capped at 0.97 (proven trust locked false);
genuine ABSTAIN examples where history is insufficient — an absent value carries no value;
fixed caveats in every answer: paper-only, advisory, not financial advice.

Composition (2,693 rows): ENTER_LONG 1,267 · EXIT_LONG 773 · HOLD 637 (downsampled 1-in-7 per stream, declared rule) · ABSTAIN 16. Assets (rows): BTC 682 · ETH 707 · SOL 741 · BONK 563 (365d daily, ending 2026-07-21).

Timestamps: each row's asOfIso is the decision bar close; the engine books fills at the next bar close. Evidence windows end at the decision bar — no lookahead.

Row format

{
  "messages": [
    {"role": "system", "content": "You are SZL-Quant, a doctrine-governed advisory research analyst. LAW: …"},
    {"role": "user", "content": "{\"task\":\"advisory-signal-decision\",\"asset\":{\"coinId\":\"bitcoin\",\"symbol\":\"BTC\"},\"asOfIso\":…,\"params\":…,\"evidence\":{…}}"},
    {"role": "assistant", "content": "{\"action\":\"HOLD\",\"components\":[…],\"conviction\":null,…,\"caveats\":[…]}"}
  ],
  "provenance": {
    "derivation": "deterministic-replay",
    "sourceReceipt": "receipts/backtest_BTC_365d.receipt.json",
    "receiptSha256": "…64 hex…",
    "datasetSha256": "…64 hex…",
    "barIndex": 123,
    "generator": "tools/sft-export.mjs v1"
  }
}

Verify before you trust

git clone https://github.com/szl-holdings/szl-quant && cd szl-quant
npm test                      # includes sft-export verification suite
sha256sum sft/quant_sft_v1.jsonl   # must equal manifest.jsonlSha256
node verify/verify.mjs --dir receipts   # source receipts, independently

The manifest receipt (quant_sft_v1.manifest.receipt.json) is an in-toto/DSSE envelope signed by the szl-quant engine key (keyId 5c6cf59741ade920, pubkey committed in the repo); its subject digest pins the manifest, and the manifest pins the JSONL.

Honest limits

The underlying feed is REPORTED (public venue history) — the dataset inherits that trust level; backtest context is MEASURED replay, not market truth.
Decisions reflect one momentum/mean-reversion strategy family on 4 assets × 365 daily bars — narrow coverage, no claim of general market skill.
Labels teach the form of honest reasoning; they do not certify profitable trading. The source engine's own receipts state limited/negative out-of-sample results plainly.
Receipts are attestations by a keyholder, not cryptographic proof of computation.

Advisory research data. Paper-only lineage. Not financial advice.

Downloads last month: 118

Total file size:

5.84 MB

Collection including SZLHOLDINGS/szl-quant-sft-v1

SZL Holdings — Datasets & Evidence

Collection

Managed datasets, receipts, manifests, and diligence evidence. • 33 items • Updated 7 days ago • 1