@inproceedings{jayalath-etal-2025-prism,
    title = "{PRISM}: Efficient Long-Range Reasoning With Short-Context {LLM}s",
    author = "Jayalath, Dulhan  and
      Wendt, James Bradley  and
      Monath, Nicholas  and
      Tata, Sandeep  and
      Gunel, Beliz",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.emnlp-main.517/",
    pages = "10207--10229",
    ISBN = "979-8-89176-332-6",
    abstract = "Long-range tasks demand reasoning over long inputs. However, existing solutions are limited, e.g., long-context models require large compute budgets, parameter-efficient fine-tuning (PEFT) needs training data, and retrieval-augmented generation (RAG) entails complex task-specific designs. Though in-context approaches overcome many of these issues, methods with short-context LLMs are inefficient, trading context for processing more tokens. We introduce **PRISM**, a highly token-efficient in-context method based on structured schemas that outperforms baselines on diverse tasks with **4x shorter contexts**. This approach produces concise outputs and efficiently leverages key-value (KV) caches to **reduce costs by up to 54{\%}**. PRISM scales down to tiny contexts without increasing costs or sacrificing quality, and generalizes to new tasks with minimal effort by generating schemas from task descriptions."
}