I belive it is a hardware issue. I am having a similar issue with an Asus RTX 3060 which I use for mining, where the card won't accept any core clocks and will stay at 200 Mhz. Here is my kind of workaround which at least works for my card in a mining environment for 2 years now.
1. Set the maximum possible core clock offset (unrealisticly high) regardless if the card accepts it
2. Set the fans to 0 rpm
3. Let the card run until it runs into a thermal crash at around 90 C core
4. Reboot
5. Check if the card still runs at just 200 Mhz
6. Repeat 1 - 4 until it does not anymore (Take my card 3 - 6 attempts)
It seems like the card needs some sort of pre-heating due to either some bad solder points or actual issues with the silicon, because if I only switch it off for a couple mintes to let it cool down it will start up again with normal clocks straight away, but if I let it cool down a longer time (2 hours+) e.g. when I replace the thermal pads or give it a regular clean, it allways falls back to the locked 200 Mhz. Another symptom my card is showing as well, while in the let's call it "locked state" it will show the current power consumption fine, but when in normal operating stated consumption reading will always be 1 Watt.
Not sure if this only works if you mine on the card or also for gaming use and especially not sure if it might damage the card in the long run, but for me personally it works fine for over 2 years now.
EDIT: I my case it has nothing to do with LHR or a general issue with that model. I have more cards which are exactly the same, which have no issues at all.