On three small harmfulness benchmarks, DPO alignment raised Llama-2-7B safe-response rates from about 40 percent to 93-100 percent while QA accuracy stayed roughly flat.
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
-
Safe to Serve: Aligning Instruction-Tuned Models for Safety and Helpfulness
On three small harmfulness benchmarks, DPO alignment raised Llama-2-7B safe-response rates from about 40 percent to 93-100 percent while QA accuracy stayed roughly flat.