The facility reported two potential dark matter sightings in 2010, but those later turned out to come from the instrument itself.
The new results - to be posted to the preprint server Arxiv, external - show three signals that should only have a 0.19% chance of showing up if there were no particle causing them, around the level of what physicists call a "three-sigma" result.
They suggest a less massive dark matter particle than commonly considered - about nine times that of the proton - but one that is still consistent with some theories.
Further, it does not match up with results published in Physical Review Letters, external from another underground experiment in Italy called Xenon.
More clarity will come, as is often the case, with more data.
"We're actually in a good situation in the whole of the field," said Bernard Sadoulet of the University of California Berkeley, referring to the four types of experiments being used around the world - and within it and below it - to finally track dark matter down.
"These four ways… we can rely on to help us work through these hints," he told the meeting.
The hint-laden data were taken when CDMS used detectors made of silicon; it has since moved on to germanium detectors.
But Prof Cabrera said the team was aiming to re-install the silicon, just to check on the unexpected hints announced here.
"The fact that we're looking at changing our strategies to include silicon means it's certainly exciting."