uyghur-headword-lists: headword inventories from three printed Uyghur dictionaries

Open, machine-readable headword lists extracted from three printed Uyghur dictionaries: 131,731 written forms and 48,182 sub-entry phrases in the Uyghur (ئۇيغۇر) script. Every entry is traceable to the scanned page column it was extracted from and carries a per-round audit status recording whether a native speaker has verified it against the original scan; 40,430 forms (30.7%) are verified so far. The lists from the 2009 imla dictionary and the 2011 explanatory dictionary follow the 2009 orthographic standard; the six-volume explanatory dictionary of the 1990s predates that standard and is published in its original spelling. Released as plain TSV under CC BY 4.0. Chinese transliterations tagged in the source dictionaries' own etymology columns were reviewed individually; 1,105 rows were removed and the full removal list is published so the decision can be audited or reversed. The release also includes a benchmark quantifying frontier-model reliability on post-OCR correction of Uyghur, covering all three dictionaries: of 15,934 machine-proposed corrections verified by a native speaker against the scans, 60.0% were correct. The model was strongest on structural errors such as skipped headwords (92–98% correct in the explanatory dictionaries) and weakest on spelling, where it pulled toward familiar forms rather than reading what was printed: replacing ۇ with و was right 4–6% of the time in both the 1990s and 2011 dictionaries, and deleting a ى from a 1990s headword was right 7%. Self-reported confidence by gemini-3.1-pro-preview carried no signal: virtually every finding, right or wrong, was labelled high. All 16,149 proposals are published with the model's explanation and the native speaker's verdict, so every figure can be recomputed. This dataset is corrected continuously. Cite the version DOI for reproducibility.