Top 5000+ Korean Words Detailed

This is a deck I created based off publicly available lists put out by the Korean government of the top ~6000 used words in the Korean language by occurrence. That said when cleaning up the dataset I found some things that were odd and maybe fragments of words (maybe showing up due to their high occurrence in other words or maybe just an error. I also found a few duplicates here and there but regardless after a little cleaning its still around 5460 words in this deck in total.

There are other occurrences of decks out there that use this list as I don't think its a secret, pretty sure it was developed originally for people studying TOPIK. That said I like the idea that the deck is comprised of frequently used words as a means to expand vocabulary in a slightly targeted fashion. I still think this deck is likely better for intermediate/advanced learners though as I purposely added a ton of information to the cards with the goal of helping myself better memorize some of the more conceptual vocabulary or things that are harder to remember at times. I think someone who lacks deeper contextual knowledge or grammar may not be the best target for this deck but it does have a lot of information in it so maybe it would be useful depending on the person.

The deck itself is structured as follows:

Front of Card: -Korean Word

Back of Card: -English Meaning -Spoken Example -Written Example -Context/Usage Notes

I wanted to include all the extra examples and context notes specifically to help remember more difficult words or to discover the situations where sometimes its rare for a word to be spoken but common to be written or vica versa. Sometimes even same word different meanings in different context, etc.

To create this deck I leveraged some custom python scripts I made with some extremely detailed prompts that I spent a few days refining until I was satisfied with the end result. From there I leveraged OpenAI API's to go through and programmatically create the context/translations through each category on recent 4o and 4.5 models (did a lot of cross checking on the final CSV). I also did hours of manual verification to ensure any entries that were missing any fields were corrected.

So if you like verbose decks with a lot of information this will be a good deck I think. I'm excited to start using it, I have been studying for over 4 years though so I am looking to get richer levels of content in the deck that i'm using. This and the other deck I created recently by processing public subtitle data were great fun to make and I am sharing here in the event any one else stumbles across them and finds it useful.

This is v1 so as I use it if I see anything strange I'll continue to update it.