This is (going to be) a blog by Ricky_Lili / Ruiqi Li.

A master candidate of CS (NLP) in Sichuan University, a potential indie developer, a lifelong learner.

Paper in ACL 2025 Student Research Workshop:

LIP-NER: Literal Patterns Benefit LLM-Based NER

Upcoming:

* A personal blog, "My Archive".

* Some ideas about applications. The first one is "Vocab Dreall", and its demo will be available very soon.

"Vocab Dreall"

= a fusion of "Vocabulary Drill" and "Read"

This may be the first "article-native" vocabulary drilling app, powered by LLMs. LLM is not the core of this concept, article is, LLM is the (temporary) way to get articles.

I am still a guy in school, it's just not feasible for me to provide LLM service in a free way. So I figured out 3 ways to let users use LLM services without taking the financial burden myself:

"Full-auto": Watch ads to use my LLM-API keys. In this way, the ad provider pays.
"Semi-auto": Let users bind their own LLM-API keys (services held by the users themselves or the third party). In this way, the users run or pay.
"Manual": Show formulated prompts, the users copy n paste the prompts to their LLM chat apps, then copy n paste the generated articles back to my app. In this way, LLM chat service providers pay.