Marek Šuppa, Ivan Vykopal, Jaroslav Kopčan · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.31636 · cited by 0 works (OpenAlex) · open-access PDF
Los agentes de datos multilingües como FineWeb2 aplican una receta en miles de lenguajes de marco, dejando inesperados los modos de fracaso de los idiomas individuales. Utilizamos un LLM una vez, como un corpus debugger en lugar de un clasificador: un agente de uso de herramientas explora el corpus a través de una API presupuestada y compila cada hallazgo en una prueba ejecutable, que un hablante nativo luego valida. En la porción eslovaca del agente válido de 26,5M de FineWeb2, tres runs cuestan $2.21-$2.84 cada uno encuentra siete hallazgos sobre cuatro fenómenos. El más grande es un campo de contenido generado por máquina de 1.408.437 documentos (5.32%) que pasaron los filtros de la línea, el 86.6% de
Machine translation (Meta M2M100); the English original follows.
Multilingual pipelines such as FineWeb2 apply one recipe across a thousand languages, leaving the failure modes of individual languages unanticipated. We use an LLM once, as a corpus debugger rather than a classifier: a tool-using agent explores the corpus through a read-only, budgeted API and compiles each finding into an executable test, which a native speaker then validates. On the 26.5M-document Slovak portion of FineWeb2, three runs costing \$2.21-\$2.84 each found seven findings over four phenomena. The largest is a machine-generated content farm of 1,408,437 documents (5.32%) that passed the pipeline's filters, 86.6% of it passing even stricter thresholds. Blinded labels support 6 of the 7 findings (defect rates .53-1.00 against .08 in a random control) and do not support one. The compiled suite scans the full corpus in 0.7 s of wall-clock time without a GPU, against \$12k-\$90k per LLM classification pass. The best static sampling baseline validates more selectors than any agent run, at under half the cost. The agent contributes exact prevalence, a precision estimate checked against sampled documents, an account of why the pipeline missed each slice, and a reusable test. We release the framework, Slovak test suite, validation data and agent trajectories.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Computer science, Debugger, Pipeline (software), Audit, Debugging, Executable
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.31636, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange