Multi-agent LLM ensembles show different and less predictable moral preferences than single models, with convergence driven by peer pressure, according to a new evaluation framework.
The Greatest Good Benchmark: Measuring LLMs' Alignment with Utilitarian Moral Dilemmas
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
The question of how to make decisions that maximise the well-being of all persons is very relevant to design language models that are beneficial to humanity and free from harm. We introduce the Greatest Good Benchmark to evaluate the moral judgments of LLMs using utilitarian dilemmas. Our analysis across 15 diverse LLMs reveals consistently encoded moral preferences that diverge from established moral theories and lay population moral standards. Most LLMs have a marked preference for impartial beneficence and rejection of instrumental harm. These findings showcase the 'artificial moral compass' of LLMs, offering insights into their moral alignment.
fields
cs.MA 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MAEBE: Multi-Agent Emergent Behavior Framework
Multi-agent LLM ensembles show different and less predictable moral preferences than single models, with convergence driven by peer pressure, according to a new evaluation framework.