Please, set your mind into an 802.11 station and think about what happens during the start of an 802.11 frame reception.
In the preamble there is a field called L-SIGNAL and one of the subfields is called Rate. If the rate indicates 6mbs the rest of the frame could be any of the available PHY types in the 802.11 standards.
In the 5GHz band, it could be either a non-HT, HT, VHT, HE SU, HE MU, MU Trigger-based, or HE ER frame format.
How will you differentiate between those types of frame formats, and receive and process it correctly?

No clue, read this article

In the 802.11 5Ghz band there are several PHY types of frames. Those are non-HT/OFDM, HT, VHT, and HE. All those have different frame formats, especially in the preamble (physical header).
This could be like this:

The preamble for all PHY types starts with an equal legacy preamble for backward compatibility, and then each PHYs have its own HT, VHT, or HE preamble. The non-HT frame type goes directly from the legacy preamble to the data field.

The interesting part is that there is, for non-HT, HT, and VHT, no information bits who tell which type of PHYs the frame is. The information bits in the HT, VHT, or HE preamble is just information to the receiver on how it should process this type frame

So how does a receiver know what type of frame it receives when the Rate is set to 6mbs in the legacy Signal field?

The solution for non-HT, HT, and VHT is how the modulation is done in the two first symbols after the legacy preamble.

Some basics
All PHY types use the lowest and most robust modulation and coding scheme in the preamble, BPSK modulation, and 1/2 coding. BPSK transfer one information bit per subcarrier and the constellation map looks like this:

If we phase shift this constellation map 90 degrees we get what we call QBPSK (quadrature binary phase-shift keying). If we compare BPSK and QBPSK constellation map it looks like this (BPSK to the left and QBPSK to the right):

The algorithm is as follows (from The Next Generation Wireless LANS, by Perahia and Stacey): For each symbol that is processed, compute a metric based on the power of the real part of the data averaged over the 48 subcarriers minus the power of the imaginary part of the data averaged over the 48 data subcarriers.

A symbol modulated with BPSK will result in a large positive number, while a symbol modulated with QBPSK will result in a large negative number

non-HT (OFDM/802.11a)
When the frame type is non-HT, the data field is the next after the legacy preamble. The Rate subfield in the legacy preamble Signal field indicate either 6, 9, 12, 18, 24, 36, 48, or 54 mbs.
The next two symbols will therefore either compute a large positive number for BPSK or 0 for QPSK and QAM.

HT (802.11n/HT mixed format)
For HT frame type, the two first symbols in the HT preamble are modulated with QBPSK. And both give a large negative number.
The two first symbols of the HT preamble is the HT-SIG field consisting of HT-SIG1 and HT-SIG2

VHT (802.11ac)
For VHT frame types, the first symbol in the VHT preamble is BPSK modulated and the second symbol is QBPSK modulated. Those two symbols in the VHT preamble is the VHT-SIG-A1 and VHT-SIG-A2 field.

Summarizing non-HT, HT or VHT
The first symbol after the L-SIG in the legacy preamble is either the data field for non-HT, HT-SIG for HT or VHT-SIG-A1 for VHT frame formats.
If the receiver detects QPSK or QAM-modulation, it knows it is a non-HT frame format. If the receiver detects QBPSK it knows it is HT frame format.
If it detects BPSK it could be either non-HT or VHT. However, if the receiver detects QBPSK for the second symbol after the legacy preamble it knows it is VHT frame format

HE (802.11ax)
For HE they had to find another method, both for differentiating the different HE frame formats and for backward compatibility.

First, the preamble for HE frame formats is using the same setup as the legacy non-HT/HT/VHT PHY frame formats. Which means it uses BPSK 1/2 modulation and coding scheme and the same subcarriers setup. It is at the end of the HE preamble it changes to the new HE subcarrier setup. The frame format could be like this:

The L-Signal field from the legacy preamble is repeated as RL-SIG (repeated L-SIG). And next, after the RL-SIG comes HE-SIG-A. Since there are four different HE frame formats, the contents of HE-SIG-A are dependent of the frame type. To differentiate between those frame types it is being done as follows:

L-SIG and RL-SIG
It starts with the construction of the L-SIG in the legacy preamble. In the L-SIG the Length parameters is modified so that the receiver can do a mod3 calculation.
If LENGTH mod3 = 2, it is a HE Trigger-based frame or a HE SU frame
If LENGTH mod3 = 1, it is a HE MU frame or a HE ER frame

The L-SIG is repeated as RL-SIG. When HE stations recognize the RL-SIG, it understands it is HE frame format.

So at this moment, HE stations understand it is HE frame format and either HE Trigger-based/HE SU or HE MU/ER frame format

HE-SIG-A
After the RL-SIG is the HE-SIG-A. The station uses information in the first bit of the HE-SIG-A, together with the information from L-SIG/RL-SIG, to differentiate further.

HE SU or HE Trigger-based frame format
If LENGTH mod3 = 2, from the L-SIG/RL-SIG, check the first bit of HE-SIG-A. A “1” indicate a HE SU frame format, while a “0” indicate a HE Trigger-based frame format

HE MU or HE ER frame format
If LENGTH mod3= 1, from the L-SIG/RL-SIG, two things happen. If the first symbol of HE-SIG-A is QBPSK modulated, the frame is a HE MU frame format. If it is BPSK modulated, it is HE ER frame format. Now the first bit is set to “1”

To summarize all this information we can use this table:
Remark April. 3, 2021. There are some minor flaws in this table for the MU frames. I will update those soon

	L-SIG	First symbol after L-SIG	Second symbol after L-SIG
non-HT	BPSK	Data field BPSK, QPSK, QAM Large positive number or 0	Data field BPSK, QPSK, QAM Large positive number or 0
HT	BPSK	HT-SIG1 QBPSK Large negative number	HT-SIG2 QBPSK Large negative number
VHT	BPSK	VHT-SIG-A1 BPSK Large positive number	VHT-SIG-A2 QBPSK Large negative number
HE SU	BPSK Length mod 3 = 2	RL-SIG BPSK Length mod 3 = 2	HE-SIG-A1 BPSK First bit = 1
HE Trigger based	BPSK Length mod 3 = 2	RL-SIG BPSK Length mod 3 = 2	HE-SIG-A1 BPSK First bit = 0
HE MU	BPSK Length mod 3 = 1	RL-SIG BPSK Length mod 3 = 1	HE-SIG-A1 QBPSK
HE ER	BPSK Length mod 3 = 2	RL-SIG BPSK Length mod 3 = 1	HE-SIG-A1 BPSK First bit = 0

But non-HT and HE looks very similar
When comparing non-HT and HE we see there is none specific indication to differentiate between those frame types except HE MU, which uses QBPSK in the second symbol of the HE preamble.
A HE station will understand it is a HE frame if it recognize the RL-SIG, but there is no good indication for non-HT station to understand it is a HE frame format it shall, an will, not understand later.
The non-HT station expects to receive after L-SIG, the Service field and the start of the 802.11 MAC header, but since the HE preamble is sent with the same modulation as 6mbs non-HT (BPSK 1/2) it will fully be able to demodulate the first part of the HE preamble.
If the non-HT station fails because of bad detection of the Service field or it fails when the HE preamble changes to the HE subcarrier format during HE-STF and HE-LTF at the end of the HE preamble I have not found a good answer on

The closing
I hope this is useful and feedbacks/comments are welcome

References
The Next Generation Wireless LANs by Eldad Perahia and Robert Stacy
802.11-2016 standard
802.11ax draft 4.2
Gjermunds PHY cheat sheet, see here