Dataset Card for “Calc-gsm8k”
Summary
This dataset is an instance of gsm8k dataset, converted to a simple html-like language that can be easily parsed (e.g. by BeautifulSoup). The data contains 3 types of tags:
- gadget: A tag whose content is intended to be evaluated by calling an external tool (sympy-based calculator in this case)
- output: An output of the external tool
- result: The final answer of the mathematical problem (a number)
Supported Tasks
The dataset is intended for training Chain-of-Thought reasoning models able to use external tools to enhance the factuality of their responses. This dataset presents in-context scenarios where models can out-source the computations in the reasoning chain to a calculator.
Construction Process
The answers in the original dataset was in in a structured but non-standard format. So, the answers were parsed, all arithmetical expressions were evaluated using a sympy-based calculator, the outputs were checked to be consistent with the intermediate results and finally exported into a simple html-like language that BeautifulSoup can parse.
Content and Data splits
Content and splits correspond to the original gsm8k dataset. See gsm8k HF dataset and official repository for more info.
Licence
MIT, consistently with the original dataset.