Occasional hangs and a blue screen on Lenovo y700. Delamination of the chip?

All kind time of day! There is a laptop Lenovo y700 (i5-6300HQ 2.30 GHz, GTX 960m, 12 gb of ddr4), bought a year ago. Just two weeks ago faced a problem: sudden hang/blue screen, sometimes flavored sound “trrr”, if at this time a sound played. From time to time, there will be a reboot because the driver “stopped responding and was recovered” that is accompanied by a short darkening of the screen (well, it's a classic). Just before hovering the cursor begins to move as if with a huge ping, and then all.

To make it clearer, before the first freeze was not update windose update is available for the driver, of course, was not done overclocking (and to do something is not), nothing unusual. I wasn't even in the game, just sitting in the browser (by the way, while wrote the message, there was one freeze, again, just in the browser).

Noticed that during the low load the probability of unresponsiveness is the biggest (watching videos, surfing, just desktop without any activity), if you play the game, it was mostly all okay, even a three-hour stream is normal, and it's a huge load (relatively). Only once hung heavy in the game, the coolers were working like mad, however, they are under load always work. Actually what is the whole point that under load, the device operates more stably than without it.

Also, there are no artifacts, performance is high, as always, absolutely nothing has changed in the laptop, in addition to the appeared freezes.

Err on the GTX 960, as in the event log (I have Windows 10) quite a lot of errors “display Driver nvlddmkm sys has stopped responding and has successfully recovered” (code 4101).
So in simple noticed slight overheating - 35-45C, but “it's not exactly” since the monitoring program to pull the data for this video card work with big brakes (tried MSI Afterburner, SpeedFan, Open hardware, everywhere one and the same).
Did a driver rollback, and a complete reinstallation with deleting, gave no results, updated drivers, disconnected from the PC, almost all USB devices (all of a sudden the type of PSU can not cope, because there is one monitor, but I'm an even bigger ignoramus, so not sure).
Tried to carry out stress tests in Hades, like anything special, but the standard utilite consistently catch the hang (even when doing a test of embedded video).
Checked for dust but the laptop clean, always keep him “in limbo” to the air walked near the window, trying not to litter.

To reinstall vindovs does not see much sense, because sit on it since April, and probably it will do nothing (or think?).

Having read the terrible articles from the Internet, began to suspect delamination of the chip, but “the Canon” this action sort often accompanied by artifacts, and they, I repeat not, and never did.
Sunday plan to change the thermal paste, because I have no idea what is happening there now (judging from the slight overheating, is going on there currently).
I have a crazy theory that hangs it at low load due to “the sharp rise in temperature” (XS, does it ever happen at all), that is, when the video got load, began to heat up and the cooler still does not turn on at max rpm to cool it, and happens to Davison. But when the coolers are working at maximum (in serious game) - everything seems okay.

It all was totally fucked my brain, and so constantly at work there, and when you come back, in the evening, trying different options out out. As a cry of despair, decided to write here, maybe someone faced similar issue, or knows what it is.
If their strength is not daresay, all once, drag it in service, but hardly sooner than a month, as a lot of work and no time for all this.

In any case, thank you for your attention, if you become better/worse, and when I change the paste to accomplish your goal, maybe someone my “adventures” will be useful in solving or understanding the problem. Thank you!