Radioactivity is not terribly complicated, however the terms used and the actual effects are quite complicated. I will spoil the surprise ending here early... there is no standard conversion from beq to sievert...
First some definitions:
1bq = number of atoms required to reach 1 decay per second
1 sievert = 100 rem = joules of energy absorbed per 1 kg of mass
Now the horrible truth:
1 sievert = 0.01bq * radiation type factor * tissue factor * distance
The radiation type factor, radiation energy per emission, and tissue factor must be experimentally determined on a per radionuclide basis. Additionally, iodine is preferentially utilized by the thyroid, and radioactive iodine is most likely to affect that organ in lieu of others.
alpha = released helium
beta = released electron
neutron = neutron

x-ray = x-ray
gamma ray = gamma
Typical emissions from the reactors:
131I = beta
129te= beta
137Cs= beta
90Sr = beta
So typically alpha has a higher sievert conversion factor, and beta is safer... so there's that.
To further complicate matters, 1kg of total plant material has very little meaning, due to tissue uptake specificity. It's possible the leaves contain 90% of the radionuclides but the stems contain 50% of the mass... they then measure the radioactivity, pick the leaves and throw away 50% of the mass, nearly doubling the radioactivity. To even further complicate matters is brewing efficiency... maybe you only leech 50% of the radionuclides into the water, since they may intercalate cellular constituents that are not brewed into your drink... therefore you will not consume it.
The *most* valid way to determine how dangerous it is to drink tea, is to brew it, dehydrate it, then count the radioactivity by both radiation type (probably beta) and energy... then use that info to lookup conversion values to sieverts. To avoid this, since it's a ton of work, they just list a generalized limit on bq per kg material, because frankly people are lazy, it's expensive, and the technology to do this isn't really high throughput.
Here is a calculator that attempts to model these value conversions based on experimental tissue data on a per nuclide basis: http://www.radprocalculator.com/Gamma.aspx