An ensemble of two LoRA-finetuned 8-9B models gets 80.2% accuracy on Chatbot Arena preference prediction, slightly above GPT-4's 78.3%, but with no error bars or code.
Neural approaches to conversational ai
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
LoRA-LiteE: A Computationally Efficient Framework for Chatbot Preference-Tuning
An ensemble of two LoRA-finetuned 8-9B models gets 80.2% accuracy on Chatbot Arena preference prediction, slightly above GPT-4's 78.3%, but with no error bars or code.