I worked for a number of years for a firm called SAES Pure Gas, Inc., a wholly-owned subsidiary of SAES Getters, S.P.A. - why do we care?
Because SAES Pure Gas made (and still does make) some of the world's most effective UHP (Ultra-High Purity) gas purifiers. These are used in the semi-conductor fab industry, both as precursor bulk purifiers and in-line last point of control purifiers to prevent impurity breakthrough when flowing reaction gases to photolithography equipment - and where medical gases worry about purities of parts-per-million for human use, UHP gases are concerned with controlling impurity breakthrough at the parts-per-billion scale - and in their world, if there is a minor breakthrough, it will totally hose all the silicon wafers in the photolithography pipeline at the time, and contaminate the equipment, requiring that entire line to be taken offline and be meticulously cleaned and purged at the highest cleanroom specification you can imagine - like NASA's sattelite assembly facilities. This, as you can imagine, would be a very expensive big deal, hence the market for UHP Gas Purifiers.
At the level of pulling errant molecules out of vacuums, the operative system is not filters, but getters - typically barium getters. Getters have been in use in our electronics since the days of second-generation vacuum valves (tubes for my US readers) and were very broadly used in CRTs (Cathode Ray Tubes) including most consumer televisions back when the tube was king - a getter is a material which is reactive, and which will either chemically combine with the most commonly-expected impurities in a given vacuum system or which will adsorb impurity molecules into its surface, which is typically hyper-porous for this exact purpose -adsorbtion involves the adsorber material having enough available surface area to physically trap / adhere the relevant impurity molecules within that hyper-porous metallic spongiform reactant.
So - to answer Chasly directly: when the relevant vacuum capture reference sample is initially processed, the survey vessel assesses the most-commonly-found aromatic hydrocarbons which might convey odour, and formulates, onsite, a getter / adsorber combination material based on that reference sample's specific chemistry.
The getter is highly porous, and is adhered to the inner wall of your vacuum container. The rate of adsorbtion is duly noted, relative to the average aromatic hydrocarbon molecular density of the reference sample, and a loose prediction of "freshness" is printed on the canister label, after which there is no guarantee of "odour-free" nature. This means your vacuum samples have shrink and expiration dates - way better for marketing really: "warp-fresh vacuum from the perimeter of the Horsehead Nebula!".