After I stem tokens in R, I get a lot of single letter nonwords. Any advice on how to deal with those?

See more:

R

NLP

I get single letter words in my Structural Topic Model after my preprocessing steps. Particularly after I stem my tokens.

My code is as follows:

#read the csv
articletexts <- read_csv("Meta.csv")
summary(articletexts)

#create a corpus
corp = corpus(articletexts, text_field = 'title')
corp

#stopwords
sw <- stopwords("english")
sw

#tokenize and remove elements
tokens <- tokens(corp, remove_punct = T, remove_numbers = T, remove_symbols = T) %>%
 tokens_remove(sw) %>%
  tokens_wordstem() 

What I have tried:

I have tried gsub to remove all letters of the English alphabet.