Datasets:

jaagli
/

common-words-79k

configs:
  - config_name: default
    data_files:
      - split: whole
        path: data/whole-*
dataset_info:
  features:
    - name: alias
      dtype: string
    - name: frequency
      dtype: int64
    - name: sentences
      sequence: string
  splits:
    - name: whole
      num_bytes: 83865723
      num_examples: 79059
  download_size: 54972667
  dataset_size: 83865723
task_categories:
  - feature-extraction
language:
  - en
pretty_name: common-words-79k
size_categories:
  - 1K<n<10K

Dataset Description

"Common Words 79K" (common-words-79k) contains 79,059 words and phrases, along with some sentences from Wikipedia that include these words and phrases. It is derived from the following resources:

We select classes from ImageNet-21K based on two criteria: (1) each class contains over 100 available images, and (2) the class names appear at least five times in Wikipedia.
We then include words that met the second criterion from an English wordlist.
We collect word frequency data from the English Wikipedia for all the words and phrases above.

Data Instances

Example of data instance from the dataset:

{'alias': 'newborn_infant',
'frequency': 157,
'sentences': [
'It is also recited as a prayer for protection of a newborn infant.',
'The newborn infant was named Sawai Madhavrao.',
'Jocasta handed the newborn infant over to Laius.',
"Spider-Man manages to save them and rescue Lily's newborn infant from the supervillains.",
'After her newborn infant died, Alison Langdon mutilated herself while deeply depressed.',
...,
'The newborn infant was named \'Sawai\' Madhavrao ("Sawai" means "One and a Quarter").'
  ]
}

How to Use

from datasets import load_dataset

# Load the dataset
common_words = load_dataset("jaagli/common-words-79k", split="whole")

Citation

@article{li-etal-2024-vision-language,
    title = "Do Vision and Language Models Share Concepts? A Vector Space Alignment Study",
    author = "Li, Jiaang  and
      Kementchedjhieva, Yova  and
      Fierro, Constanza  and
      S{\o}gaard, Anders",
    journal = "Transactions of the Association for Computational Linguistics",
    volume = "12",
    year = "2024",
    address = "Cambridge, MA",
    publisher = "MIT Press",
    url = "https://aclanthology.org/2024.tacl-1.68/",
    doi = "10.1162/tacl_a_00698",
    pages = "1232--1249",
    abstract = "Large-scale pretrained language models (LMs) are said to {\textquotedblleft}lack the ability to connect utterances to the world{\textquotedblright} (Bender and Koller, 2020), because they do not have {\textquotedblleft}mental models of the world{\textquotedblright} (Mitchell and Krakauer, 2023). If so, one would expect LM representations to be unrelated to representations induced by vision models. We present an empirical evaluation across four families of LMs (BERT, GPT-2, OPT, and LLaMA-2) and three vision model architectures (ResNet, SegFormer, and MAE). Our experiments show that LMs partially converge towards representations isomorphic to those of vision models, subject to dispersion, polysemy, and frequency. This has important implications for both multi-modal processing and the LM understanding debate (Mitchell and Krakauer, 2023).1"
}