This dataset card presents the results of our sentiment analysis experiments on the SST-2 (Stanford Sentiment Treebank) benchmark. We fine-tuned BERT-base with various hyperparameter configurations to find the optimal setup for binary sentiment classification.