Eliminates lookahead bias in financial backtesting through a series of yearly-partitioned pretrained LLMs.
arXiv · March 13, 2026 · 2603.11838
Why it matters
Standard LLMs contaminate financial forecasts because they are trained on future data relative to the backtest. By releasing models strictly bounded by data cutoff years, practitioners can finally trust LLM-based financial strategies without fear of data leakage.
From the abstract
In financial backtesting, large language models pretrained on internet-scale data risk introducing lookahead bias that undermines their forecasting validity, as they may have already seen the true outcome during training. To address this, we present DatedGPT, a family of twelve 1.3B-parameter language models, each trained from scratch on approximately 100 billion tokens of temporally partitioned data with strict annual cutoffs spanning 2013 to 2024. We further enhance each model with instruction