Operating on compressed semantic tags instead of raw text gives more controllable and effective instruction complexity augmentation for LLM fine-tuning, with reported gains over Evol-Instruct, Tree-Instruct, Auto-Instruct-Evol, and CodecLM.
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
-
TAG-INSTRUCT: Controlled Instruction Complexity Enhancement through Structure-based Augmentation
Operating on compressed semantic tags instead of raw text gives more controllable and effective instruction complexity augmentation for LLM fine-tuning, with reported gains over Evol-Instruct, Tree-Instruct, Auto-Instruct-Evol, and CodecLM.