So, here's a conundrum for you and I'm sure we've all been there.

We do reading. We make notes. We underline and highlight. We add some linkages to other work. We then do this for other articles and books. Months later we come back to this article and completely forget what is in it, what it's about, and much of the content.

How do you avoid this?

I found this to be a huge problem and it got me seriously worrying about my mental capacity! The only thing that helped me was writing: not just note taking but trying to write what I had read into an article or chapter.

The other thing is that you can go back over articles quickly with specific questions in mind. For example, I've needed to come up with a few references later in my thesis for a specific point and although I can't remember all the articles very clearly, a quick keyword search of the pdf files reminds me which ones covered that point.

Happens to me all the time - important stuff I've had to re-read, but sometimes that's a good thing as your knowledge expands over time and you might understand the article in a different way when you're further down the road - and make sense of it in a completely new context.
Using a referencing software package has helped me enormously - as has converting everything to PDFs and having it all on my tablet :). But yeah - I've been avoiding doing the literature review for exactly this reason. Sadly as my deadline approaches I now have to bite the bullet...

Another thing that has happened to me is what I call the 'hunter-gatherer syndrome' - I get into a real frenzy finding new interesting articles, file them and never read them. Then a year later I 're-discover' them in a long-forgotten folder and realise that they were really quite relevant and it would have helped to actually read them sooner...

I think that's normal and you can't avoid it. I started to write down a couple of key points per paper and collect that in a Word file. This way you can read through that file from time to time rather than reading thousands of papers again. You have to be minimalistic though and really reduce it to few key points. It works for me (Biology) but I would assume that this is harder when you study e.g. history or philosophy.

Not sure if it is a waste of time, but I could turn the papers I read into a single page cheat sheet. Heading, author, main arguments, location of evidence, etc

Whenever I read an article, I always write it down (or most of the time, copy and paste items strategically). After scheming the paper to identify which part I want to focus more, first I write down full citation, and then copy and paste the important bits into Notes. Sometimes I copy the whole sentence to avoid losing out the context, sometimes I jot down important chunks of words. Then I do this for several papers under the same theme, and the resulting output is a summarized version of several articles. I save it as pdf and read in my tablet. I never trust myself to remember hundreds of articles. It's kind of like a cheat sheet, so whenever I need to cite them I know exactly from which paper I am referring because everything is written under the article's title headline.

Some really useful tips here. I also hate the fact that I sometimes encounter papers I've saved and read and even highlighted, yet remember nothing about them. Gonna try some of these!

What is your specific note-taking method? After trying several ways, (highlighting, copy/paste quotes in doc), I've found out that the Cornell note-taking method works best, and I have Scrivener for compiling the notes and holding annotated PDFs, and an offline binder.

For me, I read the article once, highlight some bits, then write a 200-250 word summary w/keywords (not much jargon or direct quotes, just for me) to begin the notes. After that, I paraphrase and thematise the highlighted bits into Cornell notes (always referencing the page number and keeping direct quotes for some parts,) in Scrivener. Since Cornell notes is about paraphrasing in your own words, it somehow cements the knowledge in my brain a bit better than just copy & paste Organise with labels, and print the notes into a binder for offline reference. I'm a fast reader, and I sometimes skip over non-relevant sections, but it probably takes only about 1 hour/article, once you get into the groove. It's amazing for lit reviews, as it works well with constellating the knowledge rather than 'he said/she said'.

Learning Cornell + putting cash on Scrivener is quite a bit of an investment, but it is really helpful for notes and drafting.

You can try Evernote or OneNote. I wrote my PhD thesis using a combination of Evernote, Scrivener and Citavi. I had in my library 4000 documents. I was drowning with information and struggling for knowledge. Not only I used to forget my notes, sometimes I also read my papers twice or three times without realising it. The software solved my problems and I was able to finish my thesis in 35 months.

I mentioned 35 months since I was so happy to finish the thesis before my funding stopped. My thesis was also changing to the very last moment. Have a look at Citavi. It greatly contributed to my success.

Oh I get what you mean now. I will check that out - thank you!