Liquescent notation is fundamentally a reminder of good diction. Close to the voiced consonant or auxiliary vowel for the little note (or half the unison note). When it affects the letter s, finish the syllable before the beat and be sure not to unvoice or otherwise alter the following consonant, e.g., Spiritus Domini has to come out as sdo, not sto, and diebus vitae has to be svi, not sfi. There is a tendency not to use liquescents when the following note is in unison, but there are many exceptions. If you're dealing with adiastematic neumes, it's important to understand the ambiguity of the notation, as each liquescent sign has two possible meanings: unison (also called augmentative), or either ascending or descending (also called diminutive). Your first two examples are both cephalici and notated exactly the same as each other in the adiastematic neumes.

Remember that during the time liquescents were first being notated, the process of writing music was different than in modern times. Today, a composer thinks of what he wants the music to sound like, writes it down, and then gives it to the singer. In other words, the notation is prescriptive.

The scribes of Laon and St. Gall did not compose music this way, nor did singers read music as a set of performance indications. The musical notation is an attempt to write down sounds after they had already been sung. In other words, the notation is descriptive.

How is it that modern singers of opera, art song, and musical theater can sing in a way that the words can still be understood in spite of the lack of liquescent notation in modern music? Because they have been trained to have good diction. Likewise, if your schola has good diction they can 100% ignore liquescent neumes without doing any violence to the chant.

What does interrupt the chant line is the practice of replacing an entire sung pitch with a consonant, rather than simply voicing the consonant a little before the end of the note. The frequency with which various manuscripts disagree about liquescents does not suggest different performance practices. It suggests a consistent practice that is sometimes written explicitly and sometimes not.

The example I always give is a sign I once saw outside a tunnel: “Driving on sidewalk prohibited.” It would be a huge mistake for archeologists 1,000 from now to assume that meant everywhere else driving on the sidewalk was permitted! Likewise, the absence of a liquescent should not suggest poor diction any more than its presence should suggest an exaggerated stopping of vocal tone.