HLA imputation result - position inconsistency

[catheriz]

hi,
For some of the imputed markers, the genomic positions in the markers' names are different than the position provided by the imputation result. for example:
Part of The imputation result:
CHR POS ID REF ALT
6 29910961 SNPS_A_387_29910717_exon2_A A T
6 29910962 SNPS_A_387_29910717_exon2_G A T
6 29910963 SNPS_A_387_29910717_exon2_T A T
6 29910978 SNPS_A_395_29910725_exon2_C A T
6 29910979 SNPS_A_395_29910725_exon2_G A T
6 29910389 AA_A_-15_29910359_exon1_V A T
6 29910391 AA_A_-11_29910371_exon1_S A T
6 32551533 AA_DRB1_96_32549647_exon3_H A T
6 32551534 AA_DRB1_96_32549647_exon3_HI A T
6 32551536 AA_DRB1_96_32549647_exon3_HIQ A T

According to the HLA imputation pipeline on the Michigan Imputation Server, the interpretation of the imputed markers are:
AA_[GENE][AMINO ACID POSITION][GENOMIC POSITION][EXON][RESIDUE]
SNPS_[GENE][GENE POSITION][GENOMIC POSITION]_[EXON/INTRON]

As you can see, the POS of the markers are not the same of the GENOMIC POSITION. In addition, all the amino acid markers have A/T as REF/ALT. And does the residue indicates the original amino acid residue or the amino acid after change?

Thank you!