Freud does suggest to use the latent to understand the manifest, just as it sounds. It sounds counterintuitive because the "latent" and the "manifest" are not used in the usual sense, they are not the latent and the manifest of the same material. The "latent" is the content that comes from Freud's technique of free association. It is assumed that it reveals what the "manifest", the dream content, conceals. In other words, the "latent", elicited by the psychoanalyst, is more manifest, in the colloquial sense of the word, about the unconscious than the "manifest" of the dreams. So their interpretation proceeds by translating the "manifest" terms into those uncovered in psychoanalytic sessions, the "latent" terms.
Here is from Freud's The Interpretation of Dreams (1913):
"ALL previous attempts to solve the problems of the dream have been based directly upon the manifest dream content as it is retained in the memory, and have undertaken to obtain an interpretation of the dream from this content, or, if interpretation was dispensed with, to base a judgment of the dream upon the evidence furnished by this content. We alone are in possession of new data; for us a new psychic material intervenes between the dream content and the results of our investigations: and this is the latent dream content or the dream thoughts which are obtained by our method. We develop a solution of the dream from this latter, and not from the manifest dream content. We are also confronted for the first time with a problem which has not before existed, that of examining and tracing the relations between the latent dream thoughts and the manifest dream content, and the processes through which the former have grown into the latter".
Derrida's objections to Freud are unlike those of modern psychology, which rejects Freud's entire presupposition of unconscious mind to be "uncovered". That he accepts. He rather objects to Freud interpreting the filler of this unconscious mind as "content". It is not specific to the unconscious, he objects to the same treatment of overt language on general post-structuralist grounds. To him, it is the web of extrinsic relations, the play of contrasts, his famed différance, that carries the meaning, not some intrinsic "content".