The best starting point is here:
Generally, there is no reason to enable hardware acceleration unless you need it – if your machine can transcode in software real time then using hardware acceleration will normally only be the same quality (or even worse). The main advantage is that if they are good enough for your use case then it takes load off your server, or if your server is not powerful enough to do it at all – you are offloading the decoding/encoding to specialist hardware in the graphics chip.
Also, depending on your source formats and clients, you may be direct playing all of your media, meaning the files are coming off your disk and straight over the network without Jellyfin having to decode or encode anything.
If you are doing transcoding, however, this can be a great feature. I have an Intel (slightly newer than yours) and transcoding a single DVD (MPEG2) to h.264 for playback on the browser or on a phone takes 200-250% CPU (i.e. two and a half full cores). Using VAAPI, I can transcode three different DVDs to three different clients real time in under 85% CPU.
For Intel you are going to want VAAPI or QSV. I believe VAAPI is easier to set up (and is built into the Docker) and QSV does not add too much, so a good start.
I believe your Intel HD Graphics 6000 is Broadwell, so I think that can encode H.264 in hardware.
I do agree, though, that the transcoding page and how the hardware acceleration fits in is hard to follow for new users. I had to do a fair bit of outside reading to get to my current level of understanding on what to choose in there.