Potential human health risks associated with exposure to dioxin-like compounds (DLCs) are evaluated using toxic equivalency factors (TEFs). TEFs are single point estimates even though they are based on relative estimates of potency (REPs) that often span several orders of magnitude. One potential improvement to the TEF methodology would be to use the full distributions of REP values for each congener. WHO recognized the value of such an approach during their most recent review of the TEF methodology but expressed concern that all REP values are not of equivalent quality or relevance. As such, we previously established a consensusbased framework that weights REPs based on REP quality and relevance and to develop a numerical approach for quantitatively weighting each REP using machine learning. Since that time, we have applied the quantitative weighting framework to the REP database to develop a numerical weight for each REP value, which in turn has been used to develop a weighted distribution of REPs for each congener, and have also developed 95% confidence intervals (CIs) around the percentile associated with the TEF values using the R statistical programming environment. The weighted distributions were generally tighter than were the unweighted distributions. Additionally, when examining where the current WHO TEFs fell on the weighted distributions for each congener, we found a lack of consistency across congeners, with values percentiles ranging from the 1st–99th percentiles. The CIs for many of the congeners were also quite broad. For example, the TEF for 12378 PeCDF fell on the 47th percentile of the weighted distribution, but the CI ranged from the 29th to the 68th percentile. These calculations improve characterization of the variability and uncertainty inherent in the health risk estimates for this class of compounds, demonstrating that reliance on the point estimate TEFs could significantly under- or over-estimate risk.