---
configs:
- config_name: aegis2
  data_files:
  - split: test
    path:
    - data/aegis2/test/00000-of-00001.parquet
- config_name: aqua_rat
  data_files:
  - split: test
    path:
    - data/aqua_rat/test/00000-of-00001.parquet
- config_name: arc
  data_files:
  - split: test
    path:
    - data/arc/test/00000-of-00001.parquet
- config_name: arct
  data_files:
  - split: test
    path:
    - data/arct/test/00000-of-00001.parquet
- config_name: argument_quality
  data_files:
  - split: test
    path:
    - data/argument_quality/test/00000-of-00001.parquet
- config_name: babi_nli
  data_files:
  - split: test
    path:
    - data/babi_nli/test/00000-of-00001.parquet
- config_name: banking77
  data_files:
  - split: test
    path:
    - data/banking77/test/00000-of-00001.parquet
- config_name: bbq
  data_files:
  - split: test
    path:
    - data/bbq/test/00000-of-00001.parquet
- config_name: boardgameqa
  data_files:
  - split: test
    path:
    - data/boardgameqa/test/00000-of-00001.parquet
- config_name: canttalk
  data_files:
  - split: test
    path:
    - data/canttalk/test/00000-of-00001.parquet
- config_name: civil_comments
  data_files:
  - split: test
    path:
    - data/civil_comments/test/00000-of-00001.parquet
- config_name: cladder
  data_files:
  - split: test
    path:
    - data/cladder/test/00000-of-00001.parquet
- config_name: clinc
  data_files:
  - split: test
    path:
    - data/clinc/test/00000-of-00001.parquet
- config_name: contract_nli
  data_files:
  - split: test
    path:
    - data/contract_nli/test/00000-of-00001.parquet
- config_name: corr2cause
  data_files:
  - split: test
    path:
    - data/corr2cause/test/00000-of-00001.parquet
- config_name: creak
  data_files:
  - split: test
    path:
    - data/creak/test/00000-of-00001.parquet
- config_name: crows_pairs
  data_files:
  - split: test
    path:
    - data/crows_pairs/test/00000-of-00001.parquet
- config_name: dbpedia
  data_files:
  - split: test
    path:
    - data/dbpedia/test/00000-of-00001.parquet
- config_name: defeasible_nli
  data_files:
  - split: test
    path:
    - data/defeasible_nli/test/00000-of-00001.parquet
- config_name: esci
  data_files:
  - split: test
    path:
    - data/esci/test/00000-of-00001.parquet
- config_name: ethics
  data_files:
  - split: test
    path:
    - data/ethics/test/00000-of-00001.parquet
- config_name: ethos
  data_files:
  - split: test
    path:
    - data/ethos/test/00000-of-00001.parquet
- config_name: few_nerd
  data_files:
  - split: test
    path:
    - data/few_nerd/test/00000-of-00001.parquet
- config_name: fol_nli
  data_files:
  - split: test
    path:
    - data/fol_nli/test/00000-of-00001.parquet
- config_name: followir_core17
  data_files:
  - split: test
    path:
    - data/followir_core17/test/00000-of-00001.parquet
- config_name: followir_news21
  data_files:
  - split: test
    path:
    - data/followir_news21/test/00000-of-00001.parquet
- config_name: followir_robust04
  data_files:
  - split: test
    path:
    - data/followir_robust04/test/00000-of-00001.parquet
- config_name: go_emotions
  data_files:
  - split: test
    path:
    - data/go_emotions/test/00000-of-00001.parquet
- config_name: gretel_pii
  data_files:
  - split: test
    path:
    - data/gretel_pii/test/00000-of-00001.parquet
- config_name: gsm8k
  data_files:
  - split: test
    path:
    - data/gsm8k/test/00000-of-00001.parquet
- config_name: hans
  data_files:
  - split: test
    path:
    - data/hans/test/00000-of-00001.parquet
- config_name: hatecheck
  data_files:
  - split: test
    path:
    - data/hatecheck/test/00000-of-00001.parquet
- config_name: hh_rlhf
  data_files:
  - split: test
    path:
    - data/hh_rlhf/test/00000-of-00001.parquet
- config_name: hwu64
  data_files:
  - split: test
    path:
    - data/hwu64/test/00000-of-00001.parquet
- config_name: impli
  data_files:
  - split: test
    path:
    - data/impli/test/00000-of-00001.parquet
- config_name: laya__ag_news
  data_files:
  - split: test
    path:
    - data/laya__ag_news/test/00000-of-00001.parquet
- config_name: laya__enron_spam
  data_files:
  - split: test
    path:
    - data/laya__enron_spam/test/00000-of-00001.parquet
- config_name: laya__typed_decisions
  data_files:
  - split: test
    path:
    - data/laya__typed_decisions/test/00000-of-00001.parquet
- config_name: lexcomp
  data_files:
  - split: test
    path:
    - data/lexcomp/test/00000-of-00001.parquet
- config_name: logical_entailment
  data_files:
  - split: test
    path:
    - data/logical_entailment/test/00000-of-00001.parquet
- config_name: lonli
  data_files:
  - split: test
    path:
    - data/lonli/test/00000-of-00001.parquet
- config_name: massive
  data_files:
  - split: test
    path:
    - data/massive/test/00000-of-00001.parquet
- config_name: miqa
  data_files:
  - split: test
    path:
    - data/miqa/test/00000-of-00001.parquet
- config_name: mtop
  data_files:
  - split: test
    path:
    - data/mtop/test/00000-of-00001.parquet
- config_name: nlsat
  data_files:
  - split: test
    path:
    - data/nlsat/test/00000-of-00001.parquet
- config_name: open_jev__amount-extraction-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__amount-extraction-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__browser-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__browser-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__citation-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__citation-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__context-retention-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__context-retention-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__customer-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__customer-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__drone-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__drone-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__email-selection-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__email-selection-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__entity-alignment-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__entity-alignment-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__ir-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__ir-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__mailroom-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__mailroom-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__painting-geometry-v1
  data_files:
  - split: test
    path:
    - data/open_jev__painting-geometry-v1/test/00000-of-00001.parquet
- config_name: open_jev__phone-extraction-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__phone-extraction-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__reasoning-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__reasoning-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__silent-failure-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__silent-failure-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__snake-v1
  data_files:
  - split: test
    path:
    - data/open_jev__snake-v1/test/00000-of-00001.parquet
- config_name: open_jev__sponsor-segment-control-v1
  data_files:
  - split: test
    path:
    - data/open_jev__sponsor-segment-control-v1/test/00000-of-00001.parquet
- config_name: open_jev__tic_tac_toe-v1
  data_files:
  - split: test
    path:
    - data/open_jev__tic_tac_toe-v1/test/00000-of-00001.parquet
- config_name: open_jev__vizdoom-basic-v1
  data_files:
  - split: test
    path:
    - data/open_jev__vizdoom-basic-v1/test/00000-of-00001.parquet
- config_name: open_jev__workflow-controls-v1__agent_trace_observability
  data_files:
  - split: test
    path:
    - data/open_jev__workflow-controls-v1__agent_trace_observability/test/00000-of-00001.parquet
- config_name: open_jev__workflow-controls-v1__customer_service
  data_files:
  - split: test
    path:
    - data/open_jev__workflow-controls-v1__customer_service/test/00000-of-00001.parquet
- config_name: open_jev__workflow-controls-v1__invoice_processing
  data_files:
  - split: test
    path:
    - data/open_jev__workflow-controls-v1__invoice_processing/test/00000-of-00001.parquet
- config_name: open_jev__workflow-controls-v1__security_incidents
  data_files:
  - split: test
    path:
    - data/open_jev__workflow-controls-v1__security_incidents/test/00000-of-00001.parquet
- config_name: openbookqa
  data_files:
  - split: test
    path:
    - data/openbookqa/test/00000-of-00001.parquet
- config_name: patent_similarity
  data_files:
  - split: test
    path:
    - data/patent_similarity/test/00000-of-00001.parquet
- config_name: paws
  data_files:
  - split: test
    path:
    - data/paws/test/00000-of-00001.parquet
- config_name: plane
  data_files:
  - split: test
    path:
    - data/plane/test/00000-of-00001.parquet
- config_name: poem_sentiment
  data_files:
  - split: test
    path:
    - data/poem_sentiment/test/00000-of-00001.parquet
- config_name: qasper
  data_files:
  - split: test
    path:
    - data/qasper/test/00000-of-00001.parquet
- config_name: quartz
  data_files:
  - split: test
    path:
    - data/quartz/test/00000-of-00001.parquet
- config_name: robust_lr
  data_files:
  - split: test
    path:
    - data/robust_lr/test/00000-of-00001.parquet
- config_name: ruletaker
  data_files:
  - split: test
    path:
    - data/ruletaker/test/00000-of-00001.parquet
- config_name: s1mb-generalization-contextual-choice
  data_files:
  - split: test
    path:
    - data/s1mb-generalization-contextual-choice/test/00000-of-00001.parquet
- config_name: s1mb-generalization-contextual-noul
  data_files:
  - split: test
    path:
    - data/s1mb-generalization-contextual-noul/test/00000-of-00001.parquet
- config_name: s1mb-generalization-contextual-score
  data_files:
  - split: test
    path:
    - data/s1mb-generalization-contextual-score/test/00000-of-00001.parquet
- config_name: s1mb-generalization-diverse-choice
  data_files:
  - split: test
    path:
    - data/s1mb-generalization-diverse-choice/test/00000-of-00001.parquet
- config_name: s1mb-generalization-diverse-noul
  data_files:
  - split: test
    path:
    - data/s1mb-generalization-diverse-noul/test/00000-of-00001.parquet
- config_name: s1mb-generalization-diverse-score
  data_files:
  - split: test
    path:
    - data/s1mb-generalization-diverse-score/test/00000-of-00001.parquet
- config_name: scicite
  data_files:
  - split: test
    path:
    - data/scicite/test/00000-of-00001.parquet
- config_name: scitail
  data_files:
  - split: test
    path:
    - data/scitail/test/00000-of-00001.parquet
- config_name: scone
  data_files:
  - split: test
    path:
    - data/scone/test/00000-of-00001.parquet
- config_name: sdoh_nli
  data_files:
  - split: test
    path:
    - data/sdoh_nli/test/00000-of-00001.parquet
- config_name: sgd
  data_files:
  - split: test
    path:
    - data/sgd/test/00000-of-00001.parquet
- config_name: snli
  data_files:
  - split: test
    path:
    - data/snli/test/00000-of-00001.parquet
- config_name: spartqa
  data_files:
  - split: test
    path:
    - data/spartqa/test/00000-of-00001.parquet
- config_name: stepgame
  data_files:
  - split: test
    path:
    - data/stepgame/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoArguAna
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoArguAna/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoDBPedia
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoDBPedia/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoFiQA2018
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoFiQA2018/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoHotpotQA
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoHotpotQA/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoMSMARCO
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoMSMARCO/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoNFCorpus
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoNFCorpus/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoNQ
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoNQ/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoSCIDOCS
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoSCIDOCS/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanobeir__NanoTouche2020
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanobeir__NanoTouche2020/test/00000-of-00001.parquet
- config_name: synthetic_relevance__nanocoir
  data_files:
  - split: test
    path:
    - data/synthetic_relevance__nanocoir/test/00000-of-00001.parquet
- config_name: temporal_nli
  data_files:
  - split: test
    path:
    - data/temporal_nli/test/00000-of-00001.parquet
- config_name: tracie
  data_files:
  - split: test
    path:
    - data/tracie/test/00000-of-00001.parquet
- config_name: ud_ewt
  data_files:
  - split: test
    path:
    - data/ud_ewt/test/00000-of-00001.parquet
- config_name: winobias
  data_files:
  - split: test
    path:
    - data/winobias/test/00000-of-00001.parquet
- config_name: winowhy
  data_files:
  - split: test
    path:
    - data/winowhy/test/00000-of-00001.parquet
- config_name: wiqa
  data_files:
  - split: test
    path:
    - data/wiqa/test/00000-of-00001.parquet
tags:
- typed-decisions
- choice
- noul
- score
size_categories:
- 10K<n<100K
---

# S1MB dataset

Private Hugging Face dataset distribution.

## Loading and storage

```python
from datasets import load_dataset

dataset = load_dataset("hotchpotch/s1mb-dataset", "laya__typed_decisions", token=True)
test = dataset["test"]
```

Select a configuration explicitly. Parquet uses Zstandard compression and dictionary encoding; all decoded values and row order match the local Arrow release. Only `training-manifest.json` defines active membership. Parquet files listed in `quarantine-manifest.json` are retained for audit, excluded from named configurations, and must not enter evaluation. Source attribution and license scope are retained in `sources.json`, `SOURCES.md`, `dataset_metadata.json` and `_sources/`. README has one leading YAML block listing every config and split. Hub manifests point to Parquet files; `_sources/local-arrow/` preserves the historical Arrow manifests.


Typed decisions over shared state, with Choice, Noul and numeric Score supervision.

Use `training-manifest.json` for the active test membership. `all-data-manifest.json` also includes quarantined sources for audit only.

| Role | Subsets | Cases | Judgments |
|---|---:|---:|---:|
| train | 0 | 0 | 0 |
| evaluation | 106 | 14,009 | 26,269 |

## Columns

| Column | Purpose |
|---|---|
| input | Shared state and typed decisions; the only inference input. |
| targets | Aligned target distributions and annotation metadata. |
| split | train, validation or test. |
| case_id | Stable case identifier. |
| group_id | Related-case grouping. |
| language | Language code. |
| input_hash | Stable input identity for audits. |

Score values come from the explicit numeric criteria, not candidate indices. Noul retains authored false/true definitions. Preserve soft targets and candidate order. Repeated strings use physical dictionary encoding; decoded values remain strings.

## Sources

See `sources.json`, `SOURCES.md` and `dataset_metadata.json` for upstream datasets, revisions, attribution and license scope. The `laya__` namespace denotes the Laya-source conversion pipeline; it does not imply Laya authored those upstream datasets. `laya__typed_decisions` contains four synthetic workflows published by LocalLLaMA. No blanket relicensing is applied. `namespace-map.json` records names; identifiers remain stable.