Download README.md from GhostScientist/semanticwiki-data-v3: direct link, hf CLI and curl.
- Browser
- Download file 3.07 kB
-
https://huggingface.co/datasets/GhostScientist/semanticwiki-data-v3/resolve/main/README.md
- Command line
-
hf download hf://datasets/GhostScientist/semanticwiki-data-v3/README.md
-
curl -L -o README.md https://huggingface.co/datasets/GhostScientist/semanticwiki-data-v3/resolve/main/README.md
SemanticWiki Data v3
Training data for architectural-wiki generation with verifiable citations.
313 examples, each a full DeepWiki-style page generated from a real GitHub
repository, where every path:line citation was deterministically verified
against the harvested repo snapshot (file exists, lines in bounds, non-empty).
Why v3
v1/v2 (semanticwiki-data) contained synthetic two-file toy snippets, a single
category, and taught format memorization rather than architecture reasoning.
v3 uses real repositories across 7 languages and 4 documentation categories.
Pipeline
- Discovery - 100 repos via GitHub search (permissive licenses, 500-100k stars, 0.5-30 MB), multi-language quota. 97 harvested, 3 rejected (markdown-only).
- Context - line-numbered rendering of each repo's core files (large files windowed, tokenizer-measured budget <= 20k tokens).
- Teacher generation - Qwen/Qwen3-Coder-30B-A3B-Instruct (vLLM), strict
path:start-endcitation format with a one-shot example. - Verification - deterministic citation checker per page.
- Repair - one batched round: the teacher sees its own draft + the exact invalid citations and rewrites. 177 repairs run, 176 accepted.
- Gate - keep pages with validity >= 0.8: 388 generated -> 313 kept.
Stats
| Repos | 97 (python 25, typescript/javascript, go, rust, java, c) |
| Pages kept / generated | 313 / 388 |
| Mean citation validity by category | overview 0.846, deep-dive 0.803, data-flow 0.780, rationale 0.830 |
| Post-repair repairs accepted | 176 / 177 |
Schema (train.jsonl)
messages-[system, user(<START_OF_CONTEXT>...<query>...), assistant]source-org/nameof the real repocategory- architecture_overview | module_deep_dive | data_flow | design_rationalequality_score- citation validity x 100citation_stats- total/validity/invalid detailslicense,language- from the GitHub registry
Files
train.jsonl- the 313 training rowsrepo_registry.json- all 100 discovered repos with licenses (also the eval exclusion list)repo_snapshots.tar.gz(272 MB) - harvested repo snapshots, for citation re-verificationraw_pages.jsonl,repair_history.json,summary.json- full provenance
License
Dataset released under Apache-2.0. Source repositories are MIT / Apache-2.0 /
BSD-3-Clause (see repo_registry.json); code contexts are excerpts under those
licenses, generated pages are teacher-model text about that code.
Usage
Train with TRL SFTTrainer using the conversational prompt-completion format
(prompt = messages[:-1], completion = [assistant]); loss is computed on the
completion only. See GhostScientist/semanticwiki-dataset-v3-scripts/sw_train.py.
Provenance
Generated with the pipeline in GhostScientist/semanticwiki-dataset-v3-scripts; evaluated in GhostScientist/semanticwiki-eval-v3.