ArticleProceedings of the National Academy of Sciences of the United States of America2023
Systematic testing of three Language Models reveals low language accuracy, absence of response stability, and a yes-response bias.
Article in Proceedings of the National Academy of Sciences of the United States of America, 2023. The graph could read no effect estimate from its abstract, so it casts no vote on the map. Cited by 21 papers.
What it found
Each row is one number read from the abstract, on the scale the paper reported it, with its interval. Left of the dashed line favours the treatment, right favours the comparator. Under each row is the sentence it came from. New to these charts? A ten-minute tutorial.
The abstract states no effect estimate the extractor could read, or names no intervention and outcome on the map, so this paper lights no cell and moves no belief. It is still indexed, cited and linked below.
The trial behind it
Trials whose registry record cites this paper, or whose number appears in the abstract. A trial that started after this paper was published is citing it as background, not reporting it.
Neither the registry nor the abstract names a trial number. If this is a trial report, that itself is worth knowing.
Who cites it
21 citing papers in PubMed.
- Large language models improve clinical decision making of medical students through patient simulation and structured feedback: a randomized controlled trial.BMC medical education · 2024Trial
- Review
- AI agents can coordinate via majority-following beyond human scale.Science advances · 2026Article
- AI-generated familiarity estimates are a useful new source of information about word knowledge in Simplified Chinese.Behavior research methods · 2026Article
- Artificial stupidity or logimorphism? How misuse of language warps our thinking about 'artificial intelligence'.European heart journal. Digital health · 2026Article
- LLMs displaying less cognitive bias are not necessarily better decision makers.Nature machine intelligence · 2026Article
- Review
- Community size rather than grammatical complexity better predicts Large Language Model accuracy in a novel Wug Test.PloS one · 2026Article
- LLM-as-a-judge for infection prevention and control and antimicrobial resistance impact: comparing three main LLMs vs. human experts' assessment.Frontiers in public health · 2026Article
- Large Language Models as Distributional Baselines for Language Tasks.Open mind : discoveries in cognitive science · 2026Article
- Collaborative intelligence in AI: Evaluating the performance of a council of AIs on the USMLE.PLOS digital health · 2025Article
- Derivational morphology reveals analogical generalization in large language models.Proceedings of the National Academy of Sciences of the United States of America · 2025Article
- Can Large Language Models Generate Useful Linguistic Corpora?: A Case Study of the Word Frequency Effect in Young German Readers.Open mind : discoveries in cognitive science · 2025Article
- Article
- Large language models display human-like social desirability biases in Big Five personality surveys.PNAS nexus · 2024Article
- Testing AI on language comprehension tasks reveals insensitivity to underlying meaning.Scientific reports · 2024Article
- Reply to Hu et al.: Applying different evaluation standards to humans vs. Large Language Models overestimates AI performance.Proceedings of the National Academy of Sciences of the United States of America · 2024Article
- Language models align with human judgments on key grammatical constructions.Proceedings of the National Academy of Sciences of the United States of America · 2024Article
- Can Generative AI improve social science?Proceedings of the National Academy of Sciences of the United States of America · 2024Article
- The sociolinguistic foundations of language modeling.Frontiers in artificial intelligence · 2024Review
Corrections and comments
PubMed lists nothing against this paper. Absence here is not a guarantee, only a check that was made.
Authors and funding
3 authors.
Funding
Abstract
Humans are universally good in providing stable and accurate judgments about what forms part of their language and what not. Large Language Models (LMs) are claimed to possess human-like language abilities; hence, they are expected to emulate this behavior by providing both stable and accurate answers, when asked whether a string of words complies with or deviates from their next-word predictions. This work tests whether stability and accuracy are showcased by GPT-3/text-davinci-002, GPT-3/text-davinci-003, and ChatGPT, using a series of judgment tasks that tap on 8 linguistic phenomena: plural attraction, anaphora, center embedding, comparatives, intrusive resumption, negative polarity items, order of adjectives, and order of adverbs. For every phenomenon, 10 sentences (5 grammatical and 5 ungrammatical) are tested, each randomly repeated 10 times, totaling 800 elicited judgments per LM (total n = 2,400). Our results reveal variable above-chance accuracy in the grammatical condition, below-chance accuracy in the ungrammatical condition, a significant instability of answers across phenomena, and a yes-response bias for all the tested LMs. Furthermore, we found no evidence that repetition aids the Models to converge on a processing strategy that culminates in stable answers, either accurate or inaccurate. We demonstrate that the LMs' performance in identifying (un)grammatical word patterns is in stark contrast to what is observed in humans (n = 80, tested on the same tasks) and argue that adopting LMs as theories of human language is not motivated at their current stage of development.
Indexed as
Identifiers
What OpenQuestion holds
Registered trials
Read under generation 80e0d062 · epoch 390. Bibliography from PubMed, PubMed Central and OpenAlex; grants from NIH RePORTER; trial links from ClinicalTrials.gov; estimates, votes and beliefs from the OpenQuestion graph.