{"ID":2863048,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.00186","arxiv_id":"2510.00186","title":"Thinkquel: A Model Dedicated to Text-to-dbt Using Synthetic Data and a Span-Aware Objective","abstract":"Transforming natural-language requests into reliable, production-ready data transformations remains challenging: correctness depends on precise schema linking and warehouse-specific SQL dialects, while the strongest supervision available during training--execution success and result matching--are provided only at the sequence level. At the same time, assembling large, execution-validated corpora is costly, and token-level objectives misalign with these global signals, yielding unstable optimization and limited portability. We introduce Thinkquel, a fine-tuned model for producing robust, portable, and execution-validated database queries. Methodologies in Thinkquel integrates a novel synthetic data pipeline, TS-SQL, that leverages dbt as a portable intermediate representation with a span-aware reinforcement learning objective, and Token-Sequence GRPO (TS-GRPO), specifically designed to bridge the gap between token-level training signals and sequence-level execution rewards when finetuning LLMs. On the 500-example TS-SQL test set, Thinkquel (32B) reaches 93.2% execution success and 61.8% exact-result match with a two-stage SFT curriculum, improving over the base model by 67.2% (exec.) and 44.4% (match). In Spider (14B) experiments, TS-GRPO increases training stability and speeds convergence of the execution-match reward relative to GRPO and GSPO.","short_abstract":"Transforming natural-language requests into reliable, production-ready data transformations remains challenging: correctness depends on precise schema linking and warehouse-specific SQL dialects, while the strongest supervision available during training--execution success and result matching--are provided only at the s...","url_abs":"https://arxiv.org/abs/2510.00186","url_pdf":"https://arxiv.org/pdf/2510.00186v2","authors":"[\"Anni Li\",\"Aria Attar\",\"Paul Dong\"]","published":"2025-09-30T19:04:53Z","proceeding":"cs.AI","tasks":"[\"cs.AI\",\"cs.LG\"]","methods":"[\"Reinforcement Learning\",\"Large Language Model\"]","has_code":false}