Datasets:
Tasks:
Text2Text Generation
Languages:
English
Multilinguality:
monolingual
Size Categories:
1K<n<10K
Language Creators:
crowdsourced
Annotations Creators:
crowdsourced
Source Datasets:
original
ArXiv:
License:
Some bad data with wrong answers in this dataset
#5
by
jonasmueller
- opened
Some of the examples in this dataset have the wrong answer.
I automatically detected some of these with my company's software, and shared them here:
https://huggingface.co/datasets/Cleanlab/bad_data_gsm8k_svamp.csv
Example:
Question: After scoring 14 points, Erin now has three times more points than Sara, who scored 8. How many points did Erin have before?
Answer According to the Dataset: 18