dleemiller / WordLlama

Things you can do with the token embeddings of an LLM
MIT License
1.39k stars 47 forks source link

query autocomplete through beam decoding #43

Open VoVAllen opened 2 weeks ago

VoVAllen commented 2 weeks ago

I am looking for some solutions for query autocomplete using LLM. Is it possible to obtain beam decoding results and possibilities by modifying WordLlama and use them as autocomplete candidates?