Expert-annotated dataset for band gap prediction

<p>Here we present the dataset for &quot;Toward Accurate Interpretable Predictions of Materials Properties within Transformer Language Models&quot; (<a href="https://doi.org/10.48550/arXiv.2303.12188">arXiv:2303.12188</a>).</p> <p>The&nbsp;<strong>dataset_annotated.json</strong>&nbsp;file is organized as follows:</p> <pre><code class="language-python">{ "JARVIS-DFT id": { "text": "text description of material generated within the Robocrystallographer library", "tokens": "sequence of tokens generated by the MatBERT tokenizer", "rationales": "rationales proposed by a domain expert", "label": "label predicted by the MatBERT model", }, ...: ..., }</code></pre> <p>&nbsp;</p>