Hugging Face's logo

Datasets:
NODEMIND
/
shadow-o-tokens-20b

Tasks:
Text Generation
Modalities:
Tabular
Text
Languages:
English
Size:
< 1K
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
array_sha256
stringclasses
10 values
blocks
int64
485
977
context_length
int64
2.05k
2.05k
created_unix
int64
1.79B
1.79B
dtype
stringclasses
1 value
folder
stringclasses
1 value
schema_version
int64
1
1
source
dict
tokenizer_identity
dict
tokens
int64
993k
2M
5ddf0f20da09c720549c10f3f50949def5d5516c4de07879b6b21fc0727e6bdd
977
2,048
1,785,129,500
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
3b0ffe964dab170e919ac7847129d7cb244af54ac59630972a59c813857dfb15
977
2,048
1,785,129,513
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
653b2698e9869b4b9ede812c23892d99a0abbdcf2942e3bca8dd3853ee71b55b
977
2,048
1,785,129,524
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
1eec2e5ed00128185a681021e4a17deab76df21e9ace17591dd6d6b63a9b6f7a
977
2,048
1,785,129,534
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
7b28c9793e99aff02c5812c181ec741abb2dc37f0c14ba42f91b8426f612bc33
977
2,048
1,785,129,544
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
ed110656b337e9540775d62d45e6f8daa1843151cf32f33fe6db2f377e147426
977
2,048
1,785,129,554
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
fa6f0cd0a69bce3a5d8616fc047119540cf1dd2bf6d70c8cf18cb12cd50436d4
977
2,048
1,785,129,565
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
d919117f2be741d972286d15c596e0e3ee49763661df087958d32bdc3f645af9
977
2,048
1,785,129,575
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
f85ef4c285e3aaf6d84791c35b3d60d9878279a8277248f10fc444fad3073a15
977
2,048
1,785,129,584
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
2,000,896
0ec514716b6d13ca02881cb251aa99a1f6b29e862a0bc767d9b75721a4ecce40
485
2,048
1,785,129,595
uint32
stage_4_anneal/benchmark-train-unique
1
{ "datasets": [ "openai/gsm8k:main:train", "allenai/ai2_arc:ARC-Challenge:train", "hotpotqa/hotpot_qa:fullwiki:train", "cais/mmlu:all:auxiliary_train", "google-research-datasets/mbpp:full:train" ], "kind": "benchmark_train_only", "sealed_eval_splits": [ "openai/gsm8k:main:test", "all...
{ "id": "allenai/OLMo-2-0425-1B", "revision": "a1847dff35000b4271fa70afc5db10fd29fedbdf", "tokenizer_json_sha256": "73fd5254624f39a88e3faac6a8e11300fc3c735ed37880d4f4f08db898eaecca", "vocab_size": 100278 }
993,280

SHADOW-O OLMo-2 top-16 logits

Sparse next-token teacher logits for training the 252.8M SHADOW-O student.

  • Teacher: allenai/OLMo-2-0425-1B
  • Token ID space: AI2 Dolma/OLMo, first 100,278 IDs
  • Context length: 2,048
  • Stored per token: input uint32, top-16 IDs uint32, top-16 logits float16
  • Planned size: 20B tokens, approximately 2.0 TB
  • Capture code: 10_distill/capture_logits.py in the SHADOW-O workspace

Each logical shard contains three files with the same base name:

  • *.tok.npy: input tokens, shape [N]
  • *.ids.npy: teacher top-token IDs, shape [N, 16]
  • *.val.npy: unnormalized teacher logits, shape [N, 16]

The capture is resumable and commits a source shard only after all arrays have been written and uploaded. The student renormalizes teacher and student distributions across the captured top-16 set.

Planned 20B mixture

  • Foundation: 6.4B education-oriented crawl, Wikipedia, and Stack-Edu
  • Density: 4.55B mathematics, code, and reading comprehension
  • Reasoning: 4.55B mathematics, code, synthetic QA, and meta-reasoning
  • Anneal: 4.5B reasoning traces, FLAN/Tülu SFT, and verifiable problems

The underlying token streams come from AI2's public OLMo/Dolma data distributions and retain their applicable source licenses and attribution requirements.
Downloads last month
5,160
Number of rows:
10
Total file size:
152 GB