The mechanisms for acidosis for each of them are a bit complicated. They'll make a lot more sense if you understand the Stewart method of analyzing acid-base disturbances. But that's a whole other beast to tackle especially if you're in the midst of studying for Step 1.
To keep things very simple, the Stewart method relies on the Strong Ion Difference (SID). Strong ions are cations that fully ionize in body fluids and anions with a pKa of < 4 in physiological pH. SID = [strong cations] - [strong anions] = [Na + K + Ca + Mg + other strong cations] - [Cl + other strong anions (e.g. lactate)]. A normal SID is around 38. If the difference shrinks, think metabolic acidosis. If the difference increases, metabolic alkalosis.
1) Hyperalimentation: total parental nutrition tends to have a high chloride load. Plug that into the SID equation and the difference shrinks. Acidosis.
2) Iron tablets: this one I'm a little unsure about this one, but I'll take a swing at it. A common form of iron tablets is ferrous sulfate. In body fluids, ferrous sulfate dissociates into a strong cation, Fe3+, and strong anion, SO4-. But the Fe3+ form is short lived and it combines with H2O to form insoluble ferric hydroxide (Fe(OH)3). The strong cations are suddenly out of the picture in insoluble form and the SO4- excess makes the SID smaller.
If you want to reconcile this concept with the more familiar anion gap approach, think of SO4- as one of the unmeasured anions increasing the anion gap
3) INH: Warning: this is mostly theory. INH overdose can cause intractable seizures. Excessive seizure activity can manifest with excess lactate production, but lactate is usually cleared rapidly by the body. What INH might be doing is inhibiting lactate dehydrogenase from processing lactate to pyruvate specifically by acting on NAD. So excess uncleared lactate causes the anion gap metabolic acidosis.
4) Saline infusion: to keep things simple, it's the high chloride load which reduces the SID.
I wish there was a simpler way of explaining all this, but to make sense of all this, particularly Stewart's method, you'd need to delve into the chemistry of it all. Here's one paper that explains how it's all derived: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2702213/