Usually when dealing with interpass you are trying to limit your maximum interpass temperature, so you would want to measure on the hottest part of the heat affected zone which would be the weld itself. Depending on the configuration of the weld you could use temperature indicator crayons (temp sticks) or a digital thermometer, or in certain cases a thermocouple.
Some welding standards specify where to measure preheat temperature and where to measure interpass temperature.
I can tell you where not to measure it. Not in the weld crater as soon as you terminate the weld and not in the weld groove if you are using a temperature indicating crayon.
My standard practice, unless otherwise specified by the welding standard, is to measure the temperature one inch from the edge of the weld groove where the next weld pass will start and just before initiating the arc.
I do not use the infrared thermometers on any material other than carbon steel. Many metals such as aluminum, stainless steel, nickel alloys, etc. do not have the same emissivity as carbon steel resulting in erroneous temperature readings. In my experience I've seen errors of 150 degrees F when I double checked the temperatures with a thermocouple.
AWS D1.1 (5.6) and D1.5 (4.2.7) require minimum preheat and interpass temperature to be measured at least 3 inches from the point of welding. I can't find anything about where the maximum temperature is to be measured.
