README.md

metadata

license: gpl-3.0
pretty_name: Calculator

Arithmetic Expression Dataset

Source code and reproducible evaluation pipeline: calculator-benchmark on GitHub.

Evaluation across four models

Models: Qwen2.5-Math-1.5B, Qwen2.5-Math-7B, AceReason-Nemotron-1.1-7B, MathGLM-2B.

Weighted score:

$$\text{weighted\_score}=\frac{\sum _{s=1}^{15}(\text{mean}({\text{correct}}_s)\cdot s^2)}{\sum _{s=1}^{15}{s}^2}\backslash text\{weighted\backslash \_score\}\; =\; \backslash frac\{\backslash sum\_\{s=1\}^\{15\}\; (\backslash text\{mean\}(\backslash text\{correct\}\_s)\; \backslash cdot\; s^2)\}\{\backslash sum\_\{s=1\}^\{15\}\; s^2\}$$

model_id	overall_acc	weighted_score
nvidia/AceReason-Nemotron-1.1-7B	0.955667	0.912847
Qwen/Qwen2.5-Math-7B-Instruct	0.803667	0.651044
Qwen/Qwen2.5-Math-1.5B-Instruct	0.758333	0.571052

Per-model CSV files are attached in this dataset repository root.