Synthesizes mechanisms of LLM censorship across the model lifecycle and argues that the key issue is making moderation proportionate, accountable, pluralistic, and auditable rather than debating whether moderation should occur.
PACT: A contract-theoretic framework for pricing agentic AI services powered by large language models.arXiv preprint arXiv:2505.21286, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Understanding Censorship in Large Language Models: From Mechanisms to Governance
Synthesizes mechanisms of LLM censorship across the model lifecycle and argues that the key issue is making moderation proportionate, accountable, pluralistic, and auditable rather than debating whether moderation should occur.