> Do you think they really mean "per liter" of the actual sample or is this possibly the amount in an already diluted solution that they made from the sample? It seems like an odd unit of measurement otherwise.

The article doesn't specify. But it doesn't particularly matter either way.

Basically, don't drink the fire retardant, and you'll be fine. Even habitual exposure is not a big deal at these levels.