In simple terms, a password hashing algorithm $pH$ takes a salt $salt$ and entered password $pwd$ and calculates $h' = pH(salt,pwd)$ and check $h$ with the current stored password's hash $h$, $h \overset{?}{=} h'$.
will entering incorrect password that just hashes to the same let me in?
If you enter an arbitrary password, it may have a negligible change to have the same password hash with the salt. If this occurs this is valid and you can enter the system.
I will not call it an incorrect password, rather call another pre-image for the $h$. Apart from brute force (and rainbow tables where salt prevents this), this is the attack that needs to be performed by the attackers. This attack actually a pre-image attack;
- Given a hash value $h$ find and input such $m$ that $h = hash(m)$. The generic cost of this attack is $\mathcal{O}(2^n)$ for $n$-bit hash functions (passord hashing or cryptographic hash functions)
The weakest point in password cracking is the human factor. Humans tend to have weak passwords and currently one of the good methods is the dicewire. Use dicewire to generate your password, or better use a password manager where they generate truly random passwords and you keep your password database with your strong password generated by the dicewire.
but how does it work for preventing colliding passwords being used interchangeably?
You can't prevent it, it exists by the pigeonhole principle. Rather the password hashing algorithm by design is secure against it, even you used simply MD5. MD5 is still secure against pre-image attacks, you will have ~128-bit pre-image resistance. Of course, one needs to use modern hashing algorithms, like Argon2.
in other words, are these techniques the reason incorrect pass won't let you in or is it non-working anyway?
This is how it works, it will always work. Keep in mind that, you will have a probability of $\dfrac{1}{2^{128}}$ for a random single password that matches your password's hash ( assuming 128-bit output). If the probability of an event is $\geq \dfrac{1}{2^{100}}$ we simply say that it-is-not-going-to-happen.
Even if you fear this, use two-factor authentication systems, and actually one should use it whenever possible.
A side note: finding a collision is irrelevant in password hashing in terms of collision attack where you only need to find two arbitrary inputs $a$ and $b$ such that $hash(a) = hash(b)$. In password hashing, the attacker has the hash and tries to find one that produces the same hash,