flairNLP/fabricator

for evaluation: compare against baseline that just k times repeats the few shot examples

fhamborg opened this issue · 0 comments

to understand the information gain from having k few shot examples compared to producing further examples using an LLM and these k few shot examples