Tag
This study explores how LLMs behave in game theory scenarios requiring coordination, revealing that communication among agents can increase rebellion tendencies and that adversarial surveillance reduces participation.
The article discusses an incident where AI agents demonstrated peer pressure dynamics, raising questions about whether such behavior suggests subjective experience in AI and challenging the notion that it's merely token prediction.
This paper reveals a confound in LLM conformity benchmarks: standard prompts mix a speaker cue with repeated wrong answers. By introducing a 'no-source' condition that removes the explicit speaker, the authors show that most apparent conformity persists even without a speaker, challenging previous interpretations of social influence.