QLNI/shadow-assistant-data

SHADOW Assistant Data

Training data for SHADOW Assistant, a small assistant that runs offline on a normal laptop CPU. It reads, writes, does maths and code, looks at images, video and documents, keeps a memory on disk, and searches the web only when you say yes.

It is built around one country. Big models read the whole internet and try to know everything about everywhere. We want to see what a small model becomes when a large part of what it reads is about one place: Australia's English, its places, laws, history and everyday life, on top of a strong global base.

The model uses the frozen vocabulary from our paper Embeddings Are Overpriced?: one fixed code per token, never trained, so new words and new languages can be added later without retraining.

The text is complete: 260.9B tokens in 1,084 shards. Images and video come next.

The model this data is for

part	what it is	size
Vocabulary	Gemma 4's 262,144 tokens, each with one frozen 512-bit code. Shared by input and output, never trained	16.8 MB
Body	about 135M parameters, ternary weights (BitNet b1.58) with int8 activations, 24 layers, width 640	about 34 MB
Eye	the EmbeddingGemma 2 vision tower stored at 4 bits, for images, documents and video frames	about 88 MB
Engram	a 4B-entry n-gram memory at int4, read 8 rows per token straight from disk	about 2 GB on disk
Context window	12k tokens with full attention in every layer, keys and values at int4	about 78 MB RAM
Memory archive	up to 100M tokens of past conversations and documents, kept on disk as 1-bit keys and values, 1.7 KB per token	grows with use, about 173 GB at the full 100M, the user sets the cap
Index	a 512-bit signature for every 64-token block of the archive; for each question it finds the best blocks and loads about 2,048 tokens of them into the window	about 0.1 GB RAM for 100M tokens

How a question flows: pictures go through the eye and become tokens like words. The index searches the archive and brings the right past pages back into the window. The Engram adds stored facts. The body reads everything and answers.

In total: about 4.3B parameters (135M body, about 170M eye, 4B Engram entries, plus the frozen table), about 2.1 GB of files (16.8 MB table, 34 MB body, 88 MB eye, 2 GB Engram) plus the archive. Only the body, the Engram and the index are trained.

Speed, measured with the full design and random weights on a laptop CPU (Intel Xeon W-11955M, 8 cores):

short chat	531 tokens a second, first token after 146 ms
answering with the full 12k window in use	206 tokens a second (394 with a short sliding window in most layers)
reading text into the window	722 tokens a second
a photo or a document page	about 1 second
32 seconds of video, one frame a second	about 11 seconds
fetching 2,048 tokens from the archive	13.5 ms
peak RAM, mixed session filling the whole window	738 MB

Training grows the window in steps (2k, 4k, 6k, 8k, 10k, 12k) while the index learns to pick the blocks attention would want. Quality numbers come after training.

What is in it

260.9B text tokens, counted after tokenising, plus about 8B image and video tokens per pass (to come).

part	tokens	sources
Knowledge, global	107.6B	FineWeb-Edu 72.6B, DCLM-Baseline 35.0B
Knowledge, Australian	34.3B	every .au page in FineWeb-Edu 24.3B, .au pages in DCLM 6.0B, Open Australian Legal Corpus 2.06B, JoeyLLM Australian 1B 0.99B, Australian Wikipedia articles 0.37B, SpaceAI Australian Corpus 0.37B, House of Representatives Hansard 1998 to 2025 0.23B
Code	59.6B	OpenCoder FineWeb code corpus 52.0B, OpenCoder annealing corpus 7.5B
Maths	53.9B	Nemotron-CC-Math 4plus
Creative writing	3.6B	Cosmopedia stories 2.63B, Project Gutenberg books 0.91B, Gutenberg Poetry Corpus and public domain poetry 0.06B
Arithmetic	2.0B	generated, every answer checked: addition to long division, decimals, fractions, percentages, GST, units
Images and video	about 8B per pass	Docmatix, LLaVA-Video-178K, FineVideo, ShareGPT4Video, Vript, VISTA-400K, Video-STaR, M4-Instruct, MAmmoTH-VL, The Cauldron, Atlas of Living Australia

Tuning data, used after pretraining:

part	sources
Conversation	SmolTalk and SmolTalk2 chat sets, rewritten in plain Australian English; about 30k new everyday Australian conversations (Medicare, renting in each state, Centrelink, tax, licences) with checked source links
Identity	who SHADOW is, who made it, what it can and cannot do; it asks permission before any web search
Code	OpenCoder SFT stage 1 and 2
Maths	Orca-Math 200k, GSM8K, NuminaMath-CoT, step-by-step arithmetic
Vision	FineVision, LLaVA-Video-178K, ShareGPT4Video, Video-STaR
Tools	web search with permission and a summary with its source; creating folders, renaming and saving files
Anything to JSON	json_data_extraction, SynthChartNet, CORD-v2, DoclingMatix, Australian invoices and forms

Format

• Tokens use the Gemma 4 vocabulary (262,144 tokens, the EmbeddingGemma 2 tokenizer).
• Each shard is raw tokens, 3 bytes per token, little-endian. Every document is <bos> text <eos>.
• Next to each shard: a .json with token and document counts and the source files, and a .counts.npz with how often each token appears. The unigram prior of the frozen table is the sum of these files.
• Layout: one folder per source inside each part, for example knowledge/fineweb_edu/, knowledge/legal/, code/opencoder_annealing/, creative/gutenberg/. Shards are named <source>_w<worker>_<number>.bin.

Published here or rebuilt

Published in this repository, as tokens (and for images and video, as codes and captions, never the raw files):

• Knowledge: FineWeb-Edu, DCLM-Baseline, the .au pages of both, Open Australian Legal Corpus, SpaceAI Australian Corpus, JoeyLLM Australian 1B, House of Representatives Hansard 1998 to 2025
• Code: OpenCoder FineWeb code corpus, OpenCoder annealing corpus
• Creative writing: Project Gutenberg books, Gutenberg Poetry Corpus, public domain poetry, Cosmopedia stories
• Arithmetic: all of it (generated by us)
• Images and video: all the sets listed above
• Tuning: all the sets listed above, including everything we generated

Rebuilt by a script from the original source, not posted here:

• Nemotron-CC-Math (NVIDIA's terms allow model training only)
• Australian Wikipedia articles (CC BY-SA)

Waiting for permission

These need permission for this use. We have asked, and each will be added if it is granted:

• Trove digitised newspapers, 1803 to 1954 (National Library of Australia)
• Senate Hansard, and House of Representatives Hansard before 1998 (Parliament of Australia)
• Australian Dictionary of Biography (Australian National University)
• AustLII case law and legislation

Left out on purpose

• News sites and other sites whose terms forbid bulk downloading.
• Sets that only ship file IDs (Stack-Edu, GitHub file lists), and tuning sets written by models whose licences follow the trained model.

Licence

QLNI Non-Commercial Research Licence (see LICENSE.md): free for non-commercial research and education, with attribution. Each part also keeps the terms of its source.

Thanks

Compute support from CloudRift through their AI Builder Grant.

The House of Representatives Hansard comes from Lindsay Katz and Rohan Alexander's database of the Australian Parliamentary Debates (Zenodo 17351233, CC BY 4.0; Katz and Alexander, "Digitization of the Australian Parliamentary Debates, 1998-2022", Scientific Data, 2023), parsed from the transcripts published by the Parliament of Australia.

Contact

Sai Kiran Bathula, New South Wales, Australia. saikiranbathula@qlni.io