A token-adaptive distillation method that mixes teacher and ground-truth targets only for hard tokens yields consistent BLEU gains in LLM translation.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Self-Evolution Knowledge Distillation for LLM-based Machine Translation
A token-adaptive distillation method that mixes teacher and ground-truth targets only for hard tokens yields consistent BLEU gains in LLM translation.