Although this might be something very trivial but I am stuck here and not able to understand how this is working. (considering the functions to be pure they can only access values passed to them as input).
They can access any bindings in the same scope; since a lambda binding does not define a new scope in the same way a top level binding does, the earlier bindings are still in scope and accessible.
Consider this: a multi-parameter binding such as
\a b c -> a + b + c
is really the same thing as
\a -> (\b -> (\c -> a + b + c))
(This is, in fact, how partial function application works.) a is still in scope when b is defined, a and b are still in scope when c is defined.
What you can't do in a pure function is *change* them; all bindings are read only, whether in or inward of their scope. Inward scopes can redefine bindings, but those new bindings are unrelated to the outer ones.
Also consider this: in Haskell, a function is a binding like any other. If you couldn't refer to bindings from outer scopes, you couldn't define new functions! Well, aside from anonymous ones/lambdas, which are useful but somewhat limiting if you can't name and reuse them.