Kai-Bo Huang

OracleProto: A Reproducible Framework for Benchmarking LLM Forecasting via Knowledge Cutoffs and Temporal Masking

This work introduces OracleProto, a reproducible framework for evaluating LLM forecasting in a controlled, temporally masked environment, and retains resolved forecasting tasks as reusable evaluation resources under explicit temporal boundaries, supporting continual accumulation of forecasting datasets and systematic e...