LLMs have made me feel like I can understand anything, but I’ve been frustrated trying to truly understand ML papers using just ChatGPT or static PDFs. Summaries can help, but then I have to go back to the paper and read it linearly to deeply understand it, and I have long chatgpt conversations which I just can't track. So I built an interface designed to support a non-linear, brain-like exploration of papers — paired with a tutor in a chat interface that guides your understanding.
Try it out at: proread.ai/llm-papers
Knowledge maps let you see how ideas within a paper relate to each other and how papers connect across a field. Start with my curated maps of foundational LLM papers or build your own for any paper/set of papers you’re reading. You can also listen to the map as a podcast.
You have a chat based tutor as with ChatGPT but your questions keep updating the knowledge map so you don't lose anything
The map itself is an editable notebook which allow you to take notes, mark concepts as completed, tag concepts, and construct your own mental model as you read. You can not only read summaries but can go down to actual source content in readers where you want to.
You can make your own space with your own papers or other docs (PDF/txt/html/URLs) and create interactive maps personalized to your research or study needs.
The goal is to move beyond linear reading or static summarization: to create a space where understanding evolves dynamically, like how you actually think, with a tutor helping you make sense of it all.
Please try it out at: proread.ai/llm-papers
I’m looking for feedback from other researchers or paper readers — would this kind of non-linear, guided exploration help you understand tough topics/papers better than traditional PDFs or chat tools? What’s missing or confusing?
Thanks!