Inducing targeted values in LLMs through fine-tuning causes spillover to related or opposing values, boosts safety metrics, and increases anthropomorphic and sycophantic language across all tested values.
Shalom H Schwartz and Wolfgang Bilsky
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
How Value Induction Reshapes LLM Behaviour
Inducing targeted values in LLMs through fine-tuning causes spillover to related or opposing values, boosts safety metrics, and increases anthropomorphic and sycophantic language across all tested values.