source string | source_revision string | source_report_sha256 string | source_rows_sha256 string | tokenizer_name string | tokenizer_revision string | eos_token_id int64 | seed int64 | generation_batch_size int64 | source_conditional_entropy_nats float64 | training_window_entropy_nats float64 | evaluation_kind string | note string | tokens int64 | cache_entries int64 | tokenizer_size int64 | packing string | row_length int64 | windows_per_source_row int64 | context_length int64 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
synthetic_empirical_support_power_decay_trigram | a76a1499b44bef1df9575adf24413cc63216af4fd6b88667bf2dca23ef8ef85e | a76a1499b44bef1df9575adf24413cc63216af4fd6b88667bf2dca23ef8ef85e | 5d09fb60f085c0103e285a81d0c68795ce7e0a1a7ce43f11b2060f913ac3d8c2 | eoinf/pile_tokenizer_4096 | 155b268a49b69bac2543a3eb0acb635cef3dc4b0 | 2 | 40 | 512 | 3.571402 | 3.5765 | training_sample | Same 1025-to-257 source-row packing as historical empirical trigrams; synthetic source changes. Training-sample evaluation is not held out. | 199,999,488 | 781,248 | 4,096 | Subwindows of independent source rows; 257 tokens and 256 scored targets per window | 1,025 | 4 | 256 |
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Exact experiment data archive
Archive of exact NumPy caches, split metadata and available generator/source files. No re-sampling or train/test relabeling is performed. Split metadata records seeds, tokenizer/source revisions and packing. A training-data evaluation remains a training-data evaluation; independently sampled synthetic tests are distinct from held-out natural-language documents. dataset_manifest.json records checksums; experiment_configs/ records the consuming experiments.
- Downloads last month
- 80