@inproceedings{sundararajan-etal-2025-input,
    title = "Input Matters: Evaluating Input Structure{'}s Impact on {LLM} Summaries of Sports Play-by-Play",
    author = "Sundararajan, Barkavi  and
      Sripada, Somayajulu  and
      Reiter, Ehud",
    editor = "Flek, Lucie  and
      Narayan, Shashi  and
      Phương, L{\^e} Hồng  and
      Pei, Jiahuan",
    booktitle = "Proceedings of the 18th International Natural Language Generation Conference",
    month = oct,
    year = "2025",
    address = "Hanoi, Vietnam",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/author-page-lei-gao-usc/2025.inlg-main.46/",
    pages = "795--809",
    abstract = "A major concern when deploying LLMs in accuracy-critical domains such as sports reporting is that the generated text may not faithfully reflect the input data. We quantify how input structure affects hallucinations and other factual errors in LLM-generated summaries of NBA play-by-play data, across three formats: row-structured, JSON and unstructured. We manually annotated 3,312 factual errors across 180 game summaries produced by two models, Llama-3.1-70B and Qwen2.5-72B. Input structure has a strong effect: JSON input reduces error rates by 69{\%} for Llama and 65{\%} for Qwen compared to unstructured input, while row-structured input reduces errors by 54{\%} for Llama and 51{\%} for Qwen. A two-way repeated-measures ANOVA shows that input structure accounts for over 80{\%} of the variance in error rates, with Tukey HSD post hoc tests confirming statistically significant differences between all input formats."
}