Instruction Tuning Chronologically Consistent Language Models

Songrun He; Linying Lv; Asaf Manela; Jimmy Wu

doi:10.48550/arxiv.2510.11677

Back

Instruction Tuning Chronologically Consistent Language Models

Preprint

Open access

Instruction Tuning Chronologically Consistent Language Models

Songrun He, Linying Lv, Asaf Manela and Jimmy Wu

arXiv

11/17/2025

DOI: 10.48550/arxiv.2510.11677

Files and links (1)

url

https://doi.org/10.48550/arxiv.2510.11677View

Preprint (Author's original) This preprint has not been evaluated by subject experts through peer review. Preprints may undergo extensive changes and/or become peer-reviewed journal articles. Open Access

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.

Computer Science - Learning

Quantitative Finance - General Finance

Details

Title: Subtitle: Instruction Tuning Chronologically Consistent Language Models
Creators: Songrun He
Linying Lv
Asaf Manela
Jimmy Wu
Resource Type: Preprint
Publication Details: arXiv
DOI: 10.48550/arxiv.2510.11677
ISSN: 2331-8422
Publisher: arXiv
Language: English
Date posted: 11/17/2025
Academic Unit: Finance
Record Identifier: 9985222831502771

Metrics

1 Record Views