Antibody binding is a time-dependent process, for example with western blots I prefer to do the incubation with primary antibody o/n at 4 °C, the sensitivity is higher than the more common 1 h at 37 °C.

Have you kept your setup sterile? What about protease inhibitors?

Have you done a time course experiment? Perhaps there is slow unfolding of the target protein that exposes an epitope. Is the time-course temperature-dependent?

It would be necessary to know what was left at the bench: the antibody alone or the antibody in contact with the antigen?

In the first case I would recommend to be make sure that the RT-incubated antibody and the fresh one come exactly from the same batch of antibody, otherwise the effect could be an artefact caused by inter-batch differences with nothing to do with the RT period. It could also be that small dilution errors caused the apparent difference, even if both antibodies come from the same source. The dilutions were prepared independently for both antibodies and this can be the reason of variability. If none of that happened and the difference is real, perhaps aggregation of the antibody at RT resulted in multivalent molecules and higher sensitivity for antigen binding.

If the antibody was left at RT in contact with the antigen, all the above described situations are also possible, but you have to add the influence of antigen unfolding effects, that could over-expose certain epitopes.

My general experience is that leaving antibodies at RT for several days, for instance when travelling w/o cooling conditions, does not cause neither significant activity loses nor improvements.

It would be good to compare your results with a new bottle or antibody kept under manufacturer suggested conditions of same kind. When you leave antibodies out at room temperature most of them will have non-specific bindings which can be easily misunderstood as brighter staining.