analiza aiqbench
This study introduces TPvG, a novel framework that evaluates Large Language Models by incorporating consequence feedback into sequential moral dilemmas, moving beyond isolated single-shot assessments. The findings reveal that LLM decision-making is highly sensitive to interaction format and often diverges from human reference patterns in high-stakes scenarios. These results underscore the critical need for robust evaluation protocols to ensure AI stability before deployment in interactive environments.
TPvG: A Moral Decision Framework for Large Language Models from One-Shot to Sequential Feedback
arXiv:2608.28610v1 Announce Type: new Abstract: Existing LLM moral evaluations typically present models with isolated moral vignettes and elicit a single-shot decision, neglecting a factor known to profoundly influence human moral behavior: consequence feedback. We introduce TPvG (Text-based Pain-versus-Gain), adapted from a human moral paradigm, which embeds consequence feedback into an everyday moral dilemma of not harming others versus maximising self-gain. TPvG comprises five moral decision tasks, progressing from minimal-context one-shot choices to sequential decisions with explicit consequence feedback. Our results show that LLM moral decisions were strongly affected by decision format (one-shot versus sequential), and explicit receiver feedback produced heterogeneous effects across models. Furthermore, LLM responses to explicit receiver feedback diverged from the human reference pattern, suggesting potentially different decision processes. These findings highlight the need to evaluate whether LLM moral behavior remains stable in high-stakes interactive settings.
Abstract:Existing LLM moral evaluations typically present models with isolated moral vignettes and elicit a single-shot decision, neglecting a factor known to profoundly influence human moral behavior: consequence feedback. We introduce TPvG (Text-based Pain-versus-Gain), adapted from a human moral paradigm, which embeds consequence feedback into an everyday moral dilemma of not harming others versus maximising self-gain. TPvG comprises five moral decision tasks, progressing from minimal-context one-shot choices to sequential decisions with explicit consequence feedback. Our results show that LLM moral decisions were strongly affected by decision format (one-shot versus sequential), and explicit receiver feedback produced heterogeneous effects across models. Furthermore, LLM responses to explicit receiver feedback diverged from the human reference pattern, suggesting potentially different decision processes. These findings highlight the need to evaluate whether LLM moral behavior remains stable in high-stakes interactive settings.
Artificial Intelligence (cs.AI)
arXiv:2608.28610 [cs.AI]
(or arXiv:2608.28610v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.28610 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 227 cuvinte
Subiecte conexe
The J-Space Debate, Agent Swarms, and Pacing Frontier AI - Digital Minds Newsletter #4
LessWrong · 18 sept. 2026
The Continual Constitutional Convention: How language models revise the documents that govern them
AI Alignment Forum · 13 oct. 2026
Deconfusing Corrigibility
AI Alignment Forum · 17 oct. 2026
Why I Left Google DeepMind
LessWrong · 15 iul. 2026