Metraipolis › archive › dc_c447bc6fee38
Variable names provide priors: semantic grounding enables data-efficient numerical prediction by large language models — arXiv:2611.02637
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.02637
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T10:23:50.609Z
- SHA-256 of the archived text
ccab414eae3379d862acf74d816e11782d5df7e729564f6bd922cd045c74312e- Stable URL of this page
https://metraipolis.com/documento/dc_c447bc6fee38
How to cite
Variable names provide priors: semantic grounding enables data-efficient numerical prediction by large language models — arXiv:2611.02637 — arXiv (en), https://arxiv.org/abs/2611.02637 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c447bc6fee38 (Metraipolis archive dc_c447bc6fee38, sha256 ccab414eae3379d862acf74d816e11782d5df7e729564f6bd922cd045c74312e)
BibTeX
@misc{dcc447bc6fee38,
title = {Variable names provide priors: semantic grounding enables data-efficient numerical prediction by large language models — arXiv:2611.02637},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.02637},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c447bc6fee38; sha256 ccab414eae3379d862acf74d816e11782d5df7e729564f6bd922cd045c74312e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02637. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models (LLMs) have achieved strong performance across diverse tasks, from question answering to code generation. Recent studies have found that LLMs can perform numerical prediction on tabular datasets via in-context learning. However, it remains unclear under what conditions and how LLMs use variable names and in-context examples for numerical prediction. Here we show that semantic grounding, the match between variable names and the values they describe, allows five LLMs without task-specific training to match or outperform seven supervised machine learning (ML) models across 20 datasets from nine scientific domains, particularly when examples are scarce. Reasoning traces indicate that variable names allow LLMs to draw on domain-specific knowledge acquired during pretraining. Using controlled tasks based on physical formulas, we further show that LLMs integrate domain-specific knowledge with in-context examples in a way consistent with Bayesian updating. Specifically, when the relationship between the inputs and the target cannot be inferred from the examples alone, LLMs rely on domain-specific knowledge as the prior. Conversely, as contradicting examples accumulate, LLM predictions shift progressively towards these examples, which act as the likelihood. Our work reveals how LLMs use domain-specific knowledge in numerical prediction and positions them as data-efficient predictors for data-scarce scientific domains. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.02637 — 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 2026-11-03T10:19:49.891Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).