Fine-tuning a safety-tuned LLM on a capability task erases safety behavior more than the reverse order, and this forgetting is worse for specific groups such as Muslim people in the authors' tests.
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
-
Chained Tuning Leads to Biased Forgetting
Fine-tuning a safety-tuned LLM on a capability task erases safety behavior more than the reverse order, and this forgetting is worse for specific groups such as Muslim people in the authors' tests.