{"ID":2856503,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.11677","arxiv_id":"2510.11677","title":"Instruction Tuning Chronologically Consistent Language Models","abstract":"We introduce a family of chronologically consistent, instruction-tuned large language models to eliminate lookahead bias. Each model is trained only on data available before a clearly defined knowledge-cutoff date, ensuring strict temporal separation from any post-cutoff data. The resulting framework offers (i) a simple, conversational chat interface, (ii) fully open, fixed model weights that guarantee replicability, and (iii) a conservative lower bound on forecast accuracy, isolating the share of predictability that survives once training leakage is removed. Together, these features provide researchers with an easy-to-use generative AI tool useful for a wide range of prediction tasks that is free of lookahead bias.","short_abstract":"We introduce a family of chronologically consistent, instruction-tuned large language models to eliminate lookahead bias. Each model is trained only on data available before a clearly defined knowledge-cutoff date, ensuring strict temporal separation from any post-cutoff data. The resulting framework offers (i) a simpl...","url_abs":"https://arxiv.org/abs/2510.11677","url_pdf":"https://arxiv.org/pdf/2510.11677v2","authors":"[\"Songrun He\",\"Linying Lv\",\"Asaf Manela\",\"Jimmy Wu\"]","published":"2025-10-13T17:45:24Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"q-fin.GN\"]","methods":"[\"Language Model\"]","has_code":false}