Notes. The full table including all the upper limits are provided as supplementary material to this paper.

^aL-band continuum observations are typically reported at 1.4 GHz. However, the exact frequency associated with the flux density in the synthesized image cannot be known without knowledge of the spectral properties of the sources in the map. The central frequency of our observations is 1.68 GHz, but for a typical source with spectral index α_R = −0.75, we find 1.6 GHz to be a more representative frequency of the reported flux densities (cf. Section 3).

^b The W₁ magnitudes are taken from the ALLWISE catalogue when available.

^c The |$S_{3.6\,\mu\mathrm{m}}$| flux densities are taken from Leipski et al. (2014).

^d The luminosity at rest frame |$4400\, {\mathring{\rm A}}$| is obtained by extrapolating |$S_{3.6\,\mu\mathrm{m}}$|⁠, W₁ or M₁₄₅₀ in that order of availability (see Section 4.1).

^e We use the definition |$R=f_{5\, \rm{GHz}}/f_{\nu ,\rm{4400 \mathring{\rm A}}}$| of radio loudness (see Section 4.1 and Kellermann et al. 1989).

^f The detection of J0231−2850 is not reliable, as the image shows non-thermal-like artefacts. We therefore treat it as a non-detection and leave its confirmation to future work. The upper limit provided for this source is at 3σ significance.

^g The references to the discovery papers and the papers reporting the values of z and M₁₄₅₀.

References. (1) Wu et al. (2015), (2) Wang et al. (2016), (3) Bañados et al. (2016), (4) Fan et al. (2006), (5) Carilli et al. (2010), (6) Chehade et al. (2018), (7) McGreer et al. (2006), (8) Stern et al. (2007), (9) Willott et al. (2010a), (10) Wang et al. (2011), (11) Fan et al. (2004), (12) Decarli et al. (2018).