mlsanityData to 14 Oct 2026

What Can String Probability Tell Us About Grammaticality?

arXiv 2510.16227 ↗First posted 17 Oct 2025cs.CL cs.AIcited by 16

Not on the map. Why?

Abstract (arXiv)

What have language models (LMs) learned about grammar? This question remains hotly debated, with major ramifications for linguistic theory. However, since probability and grammaticality are distinct notions in linguistics, it is not obvious what string probabilities can reveal about an LM's underlying grammatical knowledge. We present a theoretical analysis of the relationship between grammar, meaning, and string probability, based on simple assumptions about the generative process of corpus data. Our framework makes three predictions, which we validate empirically using 280K sentence pairs in English and Chinese: (1) correlation between the probability of strings within minimal pairs, i.e., string pairs with minimal semantic differences; (2) correlation between models' and humans' deltas within minimal pairs; and (3) poor separation in probability space between unpaired grammatical and ungrammatical strings. Our analyses give theoretical grounding for using probability to learn about LMs' structural knowledge, and suggest directions for future work in LM grammatical evaluation.

Key facts. Cited by 16 papers in the map (data through 14 Oct 2026). Builds on 24 of its 53 references that are in the map. First posted on arXiv 17 Oct 2025.

Builds on

24 of 53 references.

References · 24

  1. 2024The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scalecited by 1,160
  2. 2018Neural Network Acceptability Judgmentscited by 958
  3. 2019BLiMP: The Benchmark of Linguistic Minimal Pairs for Englishcited by 564
  4. 2016Assessing the Ability of LSTMs to Learn Syntax-Sensitive Dependenciescited by 354
  5. 2023Efficient and Effective Text Encoding for Chinese LLaMA and Alpacacited by 305
  6. 2018Targeted Syntactic Evaluation of Language Modelscited by 242
  7. 2020A Systematic Assessment of Syntactic Generalization in Neural Language Modelscited by 165
  8. 2019Neural Language Models as Psycholinguistic Subjects: Representations of Syntactic Statecited by 138
  9. 2022What Artificial Neural Networks Can Tell Us About Human Language Acquisitioncited by 99
  10. 2023The Generative AI Paradox: "What It Can Create, It May Not Understand"cited by 98
  11. 2019UER: An Open-Source Toolkit for Pre-training Modelscited by 62
  12. 2024Auxiliary task demands mask the capabilities of smaller language modelscited by 54
12 more
  1. 2024Language models align with human judgments on key grammatical constructionscited by 49
  2. 2023Prompting is not a substitute for probability measurements in large language modelscited by 44
  3. 2021On the proper role of linguistically-oriented deep net analysis in linguistic theorizingcited by 43
  4. 2023Modeling rapid language learning by distilling Bayesian priors into artificial neural networkscited by 42
  5. 2022Can language models handle recursively nested grammatical structures? A case study on comparing models and humanscited by 41
  6. 2020CLUECorpus2020: A Large-scale Chinese Corpus for Pre-training Language Modelcited by 36
  7. 2022SLING: Sino Linguistic Evaluation of Large Language Modelscited by 15
  8. 2023Evaluating Neural Language Models as Cognitive Models of Language Acquisitioncited by 15
  9. 2024A Systematic Assessment of Language Models with Linguistic Minimal Pairs in Chinesecited by 13
  10. 2018What can linguistics and deep learning contribute to each other?cited by 10
  11. 2024What Goes Into a LM Acceptability Judgment? Rethinking the Impact of Frequency and Lengthcited by 10
  12. 2023The Causal Influence of Grammatical Gender on Distributional Semanticscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 4; 2026: 12.

16 papers.

Sort:

  1. 2026Information-Theoretic Storage Cost in Sentence Comprehensioncited by 1
  2. 2026Implicit Representations of Grammaticality in Language Modelscited by 1
  3. 2026Linear representations of grammaticality in neural language modelscited by 1
  4. 2025Language models as tools for investigating the distinction between possible and impossible natural languagescited by 0
  5. 2025Large language models have learned to use languagecited by 0
  6. 2025Large language models are not about natural languagecited by 0
  7. 2025Linguists should learn to love speech-based deep learning modelscited by 0
  8. 2026Are Language Models Models?cited by 0
  9. 2026Across the Levels of Analysis: Explaining Predictive Processing in Humans Requires More Than Machine-Estimated Probabilitiescited by 0
  10. 2026You Can't Fight in Here! This is BBS!cited by 0
  11. 2026Targeted Linguistic Analysis of Sign Language Models with Minimal Translation Pairscited by 0
  12. 2026Not-So-Strange Love: Language Models and Generative Linguistic Theories are More Compatible than They Appearcited by 0
4 more
  1. 2026Do Language Models Encode Knowledge of Linguistic Constraint Violations?cited by 0
  2. 2026When transformers learn "impossible" languages, what do they learn?cited by 0
  3. 2026A retrospective analysis on the use of LLMs to study infant syntax learningcited by 0
  4. 2026Assessing the Impact of Language Disparity on Multilingual Linguistic Ability in Large Language Modelscited by 0