Llama 3 generated 507,152 synthetic Vietnamese legal queries; using them for CoT-MAE pre-training and contrastive fine-tuning improved bi-encoder and ColBERT retrieval scores on TVPL and Legal Zalo 21.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Improving Vietnamese Legal Document Retrieval using Synthetic Data
Llama 3 generated 507,152 synthetic Vietnamese legal queries; using them for CoT-MAE pre-training and contrastive fine-tuning improved bi-encoder and ColBERT retrieval scores on TVPL and Legal Zalo 21.