Hugging Face's logo

Datasets:
Xnhyacinth
/
LongWorld-Worlds-State

Tasks:
Question Answering
Languages:
English
License:
Dataset card Data Studio Files
xet
 Community
Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    IndexError
Message:      list index out of range
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1859, in _prepare_split_single
                  original_shard_lengths[original_shard_id] += len(table)
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
              IndexError: list index out of range
              
              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/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1880, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

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.

text
string
[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (313596 > 262144). Running this sequence through the model will result in indexing errors
{
"status": "PASS",
"semantic_tasks": 200,
"capacity_bins": {
"131072": 139,
"262144": 61
},
"exact_token_ranges": {},
"production_eligible": false
}
[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (299710 > 262144). Running this sequence through the model will result in indexing errors
{
"status": "PASS",
"semantic_tasks": 249,
"capacity_bins": {
"131072": 163,
"262144": 86
},
"exact_token_ranges": {},
"production_eligible": false
}
[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (265133 > 262144). Running this sequence through the model will result in indexing errors
{
"status": "PASS",
"semantic_tasks": 194,
"capacity_bins": {
"131072": 124,
"262144": 70
},
"exact_token_ranges": {
"262144": 47
},
"production_eligible": false
}
[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (263244 > 262144). Running this sequence through the model will result in indexing errors
{
"status": "PASS",
"counts": {
"train": 657,
"eval": 147,
"short": 13,
"rejected_overflow": 1
},
"canonical_source_tasks": 818,
"primary_long_samples": 804,
"full_message_capacity_counts": {
"cap64k": 141,
"cap128k": 305,
"cap256k": 358
},
"exact_context_token_range_sample_counts": {
"other_long": 767,
"64k": 12,
"128k": 9,
"256k": 16
},
"max_full_message_tokens": 261954,
"max_observed_full_message_tokens": 263244,
"production_eligible": false,
"local_training_eligible": false
}
[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (332115 > 262144). Running this sequence through the model will result in indexing errors
{
"status": "PASS",
"semantic_tasks": 199,
"capacity_bins": {
"131072": 128,
"262144": 71
},
"exact_token_ranges": {},
"production_eligible": false
}
[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (280868 > 262144). Running this sequence through the model will result in indexing errors
{
"status": "PASS",
"semantic_tasks": 216,
"capacity_bins": {
"131072": 145,
"262144": 71
},
"exact_token_ranges": {},
"production_eligible": false
}
[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (283529 > 262144). Running this sequence through the model will result in indexing errors
{
"status": "PASS",
"semantic_tasks": 243,
"capacity_bins": {
"131072": 149,
"262144": 94
},
"exact_token_ranges": {},
"production_eligible": false
}
{
"status": "PASS",
"semantic_tasks": 189,
"capacity_bins": {
"131072": 72,
End of preview.

LongWorld Worlds State

Public, versioned recovery snapshot for the LongWorld worlds development branch. This repository stores candidate and local-pilot artifacts at their original scientific admission levels. It is not a production dataset release.

Latest additional snapshot: curriculum v2

  • Tag: state-2026-09-13-curriculum-v2
  • Code: f6cf9c7359f8ae8906b8cfb7a757c5192b5492ad on GitHub worlds
  • 384 multi-QA SFT rows / 6,144 QA instances; 288 train / 96 eval rows.
  • Four executable mechanisms, 64 world seed groups, three long-input bands.
  • Actual OpenAlex hierarchy included for restoring the pinned sampler.
  • Execution report and file/restore manifest.
  • Symbolic audit passes; domain-specific natural realism, universal strict dependency, model utility, framework release and production eligibility are not established.

The initial snapshot below is retained unchanged at its original tag and paths.

Initial snapshot (preserved)

  • Snapshot: 2026-09-13-worlds-v1
  • GitHub repository: Xnhyacinth/longworld
  • Branch: worlds
  • Commit: 31e824014f2fc0cdd8869b1e71151d9139610518
  • Tag: state-2026-09-13-worlds-v1
  • Integrity inventory: STATE_MANIFEST.json
Product Rows Train / eval Status
P64 training candidates 2,486 1,953 / 533 local training eligible; strict/framework/production false
P66 multidomain handoff 200 132 / 68 local candidate; strict/framework/production false
Capability-world pilot 108 72 / 36 symbolic pipeline audited; one rule family; utility unmeasured

P66 is a selected later handoff and overlaps the lineage of a subset of P64. The table must not be summed into an independent unique-row count. The capability pilot is fully simulated and separate from authentic-source P64/P66.

Layout 

snapshots/2026-09-13/
  p64_training_candidates/
  p66_multidomain_candidates/
  capability_world_pilot/
STATE_MANIFEST.json

Each product retains its native manifest, receipt, split files, and validation sidecars. STATE_MANIFEST.json binds every uploaded file by byte size and SHA-256 and records the matching GitHub state.

Resume production 

git clone https://github.com/Xnhyacinth/longworld.git
cd longworld
git checkout state-2026-09-13-worlds-v1
uv sync

Generate a new capability batch into a new directory:

uv run --no-sync python scripts/run_capability_world_pipeline.py \
  --output data/capability_pilots/<new-snapshot-id> \
  --seeds <unique-seeds> \
  --capacities 65536 131072 262144 \
  --workers 3

Do not overwrite the uploaded snapshot. New worlds must use new seed groups and new output directories. Length views are workload variants rather than new semantic topology families. Training utility, L4/L5 coverage, and production eligibility require separate evidence.

Public backup status (2026-09-30)

The original worlds and curriculum tags and payload paths are preserved. Public availability leaves the scientific admission labels above unchanged. Probe keys, credentials, training continuation states and downloaded caches are excluded.
Downloads last month
1,137
Total file size:
2.2 GB

Collection including Xnhyacinth/LongWorld-Worlds-State