While there are reports of the impact of specific interventions designed to reduce HIV drug and sexual risk behaviors, there are few reports of the impact of HIV behaviors, there are few reports of the impact of HIV interventions in a community-based sample. We report on baseline data from a sample of African American crack smokers who were about to participate in an intervention designed to reduce HIV-related risk behaviors. The majority were male (80%) single (70%) and homeless (52%). Data indicated that 29% of the sample had been in a previous HIV intervention in the past 12 months, the majority in a correctional setting or CBO program. There were few systematic demographic differences between the two groups. Those who had been in an intervention reported using male and female condoms significantly more frequently on all measures of condom use, had positive condom use outcome expectations for male condoms and higher affective and situational condom-related self efficacy beliefs. These data suggest that, at a community level, the spectrum of HIV risk-reduction programs does produce a significant improvement in condom use and related cognition, although there is a need to cover a greater proportion of the population. Previous exposure to interventions must be critical covariate in assessing the impact of future interventions. (Authors)