Social reasoning in AI traced to an extremely small set of parameters A new study reveals that the capacity for social reasoning in large language models, a trait similar to the human “theory of mind,” originates from an exceptionally small and specialized subset of...

Social reasoning in AI traced to an extremely small set of parameters #Science #ComputerScience #ArtificialIntelligence #AI #MachineLearning #SocialReasoning

'Inside you are many wolves: Using cognitive models to interpret value trade-offs in LLMs'
Cool idea about #empathy, with cringe title :-)

arxiv.org/abs/2506.20666

#AI #computation #language #ReinforcementLearning
#EmergentBehavior #LLM #Socialreasoning #alignment

LLMs are leveling up in social reasoning through multiplayer games like Among Us. Can AI master deception and teamwork? Explore the new benchmarks pushing these limits. #AI #SocialReasoning