NOVA selects instruction data by how familiar the base model is with the instruction and target response, yielding reduced hallucination rates and competitive instruction-following on several benchmarks.
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
-
Aligning Large Language Models to Follow Instructions and Hallucinate Less via Effective Data Filtering
NOVA selects instruction data by how familiar the base model is with the instruction and target response, yielding reduced hallucination rates and competitive instruction-following on several benchmarks.