You have a choice, in the Confidence tab, to either mark entire fits as "ambiguous", to mark selected parameters as "unstable", or neither. The option to mark parameters as "unstable" is a new (in 8.2) feature from Prism Labs.

If you choose the older approach to identify ambiguous fits, Prism places a tilde ("~") in front of the parameters that have super high dependencies. The parameter value is what is used to calculate the Y values for the fit curve. But with super high (>0.9999) dependency, you could get a curve that fits just as well with very different parameter values. These values that are preceded with tildes are really not useful.