Metraipolis › archive › dc_836cc075b248

Can LLMs Access their World Knowledge for Event Prediction? — arXiv:2610.19021

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19021
License (as archived)
CC0-1.0
Archived
2026-10-15T20:23:23.919Z
SHA-256 of the archived text
845fb4dc6220b7467f58df98ac18339cd8147213e40ee9aa6b509e0267c35210
Stable URL of this page
https://metraipolis.com/documento/dc_836cc075b248

How to cite

Can LLMs Access their World Knowledge for Event Prediction? — arXiv:2610.19021 — arXiv (en), https://arxiv.org/abs/2610.19021 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_836cc075b248 (Metraipolis archive dc_836cc075b248, sha256 845fb4dc6220b7467f58df98ac18339cd8147213e40ee9aa6b509e0267c35210)
BibTeX
@misc{dc836cc075b248,
  title = {Can LLMs Access their World Knowledge for Event Prediction? — arXiv:2610.19021},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19021},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_836cc075b248; sha256 845fb4dc6220b7467f58df98ac18339cd8147213e40ee9aa6b509e0267c35210},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19021. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recent work has questioned whether large language models (LLMs) can be useful for temporal event prediction. Despite vast pretrained world knowledge, LLMs often underperform simpler specialized models. We ask whether this knowledge is truly inaccessible, or whether current methods fail to present event prediction in a form LLMs can use. We argue for the latter and show that LLMs become effective event forecasters when the interface is designed around two choices. The first grounds each event in a natural-language description, and the second separates event semantics from the prediction space to prevent semantic bias from label associations. To test this, we introduce DescribeEvent, a benchmark of ~150K events across seven datasets and six domains with event-level descriptions. We find that capable LLMs substantially benefit from descriptions and symbolic labels, outperforming description-free variants and non-LLM baselines, while weaker models gain little. These results suggest that accessing LLM world knowledge for event prediction is an interface problem, and that event descriptions offer a practical way to elicit the knowledge LLMs already encode.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.19021 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T19:36:18.156Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).