Like a Glove Discussion [HTB] [HINTS]

Let’s talk about Like a Glove.

CHALLENGE DESCRIPTION
Words carry semantic information. Similar to how people can infer meaning based on a word’s context, AI can derive representations for words based on their context too! However, the kinds of meaning that a model uses may not match ours. We’ve found a pair of AIs speaking in metaphors that we can’t make any sense of! The embedding model is glove-twitter-25. Note that the flag should be fully ASCII ans starts with ‘htb{’.

Use a similar script to solve the challenge:

import gensim.downloader as api
model = api.load("glove-twitter-25", from_hf=True)
vec1 = model[word1]
vec2 = model[word2]
vec_target = model[target]

analogy_vector = vec_target + (vec2 - vec1)

result = model.similar_by_vector(analogy_vector, topn=1)

Exclusive content is now available for Like a Glove.

What is Exclusive Content?


You can still ask for help and specific hints in this thread.