Raw responses from the models, clean answers, post-processed predictions and evaluation results for each model and dataset using in the publication Fine-Tuning with Divergent Chains of Thought Boosts Reasoning Through Self-Correction in Language Models