Login / Signup
Fine-Tuning LLMs for Multi-Turn Dialogues: Optimizing Cross-Entropy Loss with KL Divergence for All Rounds of Responses.
Zeyu Teng
Yong Song
Xiaozhou Ye
Ye Ouyang
Published in:
ICMLC (2024)
Keyphrases
</>
cross entropy
kullback leibler
fine tuning
kl divergence
log likelihood
maximum likelihood
exponential family
language modeling
kullback leibler divergence
scoring function
mahalanobis distance
machine learning
evaluation metrics
information theoretic
model selection
distance measure
information retrieval systems