The dataset, from the IMMREP_2022 TCRSpecificity repository, contains paired TCR alpha and beta chain CDR3 sequences with VDJ gene information, sourced from 10x genomics and single-cell sequencing. It includes TCRs with known specificity to 17 peptides from VDJdb, serving as positive samples, and negative controls from an unpublished dataset of CD8+CD96+ cells, along with random samples from other epitopes. The data is split 80:20 into training and test sets for each epitope, with training files including full-length TCR sequences, CDR1, CDR2, and CDR3 regions for both chains, and a label indicating positive or negative status. Its purpose is to support the prediction of TCR specificities.