MEKiT improves LLM emotion-cause pair extraction by adding emotional label knowledge to instruction prompts and mixing causal examples into training data, achieving 61.49 F1 on NTCIR-13.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MEKiT: Multi-source Heterogeneous Knowledge Injection Method via Instruction Tuning for Emotion-Cause Pair Extraction
MEKiT improves LLM emotion-cause pair extraction by adding emotional label knowledge to instruction prompts and mixing causal examples into training data, achieving 61.49 F1 on NTCIR-13.