Fiction-derived normative simulacra, when used for SFT plus GRPO with contrastive scoring, improve LLMs' contextual privacy reasoning and correlation with human expectations on real-world benchmarks.
arXiv:2306.11644 [cs]
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
The Stakeholder Grounding Exercise shows neural text embeddings are 19-26pp less reliable than human experts at capturing semantic distinctions, with misalignment strongly correlated to poorer clustering performance (ρ=0.9), replicated across Danish policy and US AI domains.
citing papers explorer
-
Reinforcing privacy reasoning in LLMs via normative simulacra from fiction
Fiction-derived normative simulacra, when used for SFT plus GRPO with contrastive scoring, improve LLMs' contextual privacy reasoning and correlation with human expectations on real-world benchmarks.
-
Grounding Text Embeddings in Stakeholder Associations
The Stakeholder Grounding Exercise shows neural text embeddings are 19-26pp less reliable than human experts at capturing semantic distinctions, with misalignment strongly correlated to poorer clustering performance (ρ=0.9), replicated across Danish policy and US AI domains.