In June, I shared a Bonus Article That, along with my curated and bookmarked research paper lists for paid subscribers, is what makes this Substack possible.
Similarly, to thank all the kind supporters, I have compiled below a list of interesting research articles that I had bookmarked and categorized from July to December 2025.
I took a cursory look at the abstracts of these papers but read only a very small excerpt. However, I still like collecting these organized lists because I go back to them often when working on a given project.
By the way, I was also working on my annual LLM review article, Status of LLM 2025: Progress, Problems and PredictionsWhich I also published today. You can find it here:​
Status of LLM 2025: Progress, Progress and Predictions
Originally, I had planned to include this list in the above article. However, the article was already getting quite long, so I decided to share the list here in a separate post instead. I hope you don’t mind receiving two emails today. My thinking was that splitting things up would make both articles easier to read, scan, and revisit later without getting lost in a very long page.
The categories of this research paper list are as follows (you can use the table of contents in the web view of this article to navigate directly to them):
-
logic model
-
1A. training logic model
-
1B. guess-time reasoning strategies
-
1c. Evaluating LLM and/or understanding the logic
-
Other reinforcement learning methods for LLM
-
Other estimation-time scaling methods
-
Model Release/Technical Report
-
architecture
-
skilled training
-
Diffusion-Based Language Model
-
Multimodal and vision-language models
-
Data and pre-training dataset
