These photos are from the same observation
Curtain trap of sticky silk threads constructed by a predatory fungus gnat larva, suspended under a large leaf in the understory of lowland rainforest. Based on my reading, in South America, the species whose larvae are known to make such two-dimensional traps outside of caves are members of the genus Neoditomyia. See Coher 1996 (www.biodiversitylibrary.org/page/56138637#page/162/mode/1up) and Matile 1997 (www.online-keys.net/sciaroidea/add01/Matile_1997_Phylogeny_&_evolution_of_larval_Diet.pdf).
Found at night at eye level in lowland terra firme forest. I consulted Rob Voss (AMNH) and he thought it could be Gracilinanus emiliae, which “would be a significant range extension, but perhaps not a very surprising one” (there are records from Meta, Colombia and Loreto, Peru - see Voss et al. 2009 at https://mn.sarem.org.ar/article/on-the-diagnostic-characters-ecogeographic-distribution-and-phylogenetic-relationships-of-gracilinanus-emiliae/). Manus appears to be paraxonic (hand digit V being nearly as long as digit IV) which agrees with Gracilinanus and Marmosa (see Table 2 and Fig. 3 in Voss et al. 2004 at https://digitallibrary.amnh.org/handle/2246/2776). G. emiliae differs from congeners in its white ventral pelage (gray/brown in the remaining species) according to Voss et al. 2009 (op. cit.).