During HTML parsing, are *all* named character references replaced by their corresponding glyph?

(I apologize if this is OT for this mailing list.)

>From what I understand, named character references, e.g. &, only exist
in HTML source code, and once the source code is parsed into the DOM, *all*
named character entities are replaced by their corresponding glyphs. There
is no exception to this rule.

For instance, this source code:

<span>&amp;</span>

will produce a DOM element (of type "span") which contains a single Text
node which then in turn contains the text value "&". So, not the entity
&amp; but the actual "&" literal character.

Could you confirm that the above is correct?

-- @simevidas