This is a set of images of an extreme object at z=4.5 in the COSMOS survey. It is the most distant mm source not associated with an optically bright quasar. The rest-frame UV and Lya imaging (the first five panels on the left) show emission near the lower left of the panes, although no emission is detected in the B-band because it falls blueward of the Lyman break. At longer wavelengths, the emission shifts to the upper left; the panel on the far right shows radio contours.
The estimated star formation rate is 1000-4000 Msun/yr, based on several indicators. The authors argue against significant AGN activity -- which would mean this SFR could be a severe overestimate -- because there is no xray detection and because an optical spectrum shows no hint of an AGN. But, as the authors note, an AGN could lie outside the optical slit. In fact I think this a fairly likely explanation since the authors placed the slit on the center of the UV emission (lower left), whereas an AGN would be expected to be associated with the longer wavelengths (upper right), which is where the most of the stellar mass and radio activity is.