Separate chronology preference from input-order copying

Establish whether models’ chronological outputs in the free ordering and free prose tasks reflect an intrinsic preference for chronological narration rather than simple preservation of the chronologically ordered input, using a shuffled-input control.

Background

The free ordering and prose tasks provide events in chronological order. Although the ordering results indicate that models do not merely copy the input exactly, the prose setting makes chronological preference and input-order preservation indistinguishable.

A shuffled-input condition would allow the two explanations to be separated and would clarify what the benchmark measures about spontaneous narrative ordering.

References

In prose the two readings coincide, and separating a preference for chronology from a preference for the given order needs a shuffled-input control we have not run.

— Can LLMs Tell Stories Out of Order? A Corpus and Benchmark for Nonlinear Narrative Generation  (2610.21468 - Raiyan et al., 15 Oct 2026) in Limitations, paragraph “Input order is a confound for the free tasks”