That's an interesting question. Obviously, the IPCC assessment reports are an important source of information. The probability of a particular temperature increase will depend on the GCMs included in the assessment and whether you assign specific probabilities to the RPC scenarios. In principle, the RCP scenarios do not have a particular probability and hence it is hard to determine the likelihood of a particular temperature increase.
I assume you are interested in the likelihood of a global temperature increase? The likelihood of regional or even local temperature increases will be even more difficult to assess and may include uncertainties due to downscaling as well.
Thank you for your reply. I came across the following link that suggests the +4° scenario is 93% probable to happen (in case of business-as-usual) in 2100.
Was wondering if there are more sources. I need the data to forcast the performance of a building with passive design strategies in the future. Knowing the probable temperatures would help.