Related papers: The Ratchet Effect in Silico through Interaction-D…
While existing works have emphasized how elites shape mass opinion, we ask whether the reverse also holds: do audience reactions on social media actively shape elite behavior? We examine this question through the lens of cross-partisan…
We propose a variation of the standard genetic algorithm that incorporates social interaction between the individuals in the population. Our goal is to understand the evolutionary role of social systems and its possible application as a…
The ability to combine linguistic guidance from others with direct experience is central to human development, enabling safe and rapid learning in new environments. How do people integrate these two sources of knowledge, and how might AI…
Multi-agent learning is intrinsically harder, more unstable and unpredictable than single agent optimization. For this reason, numerous specialized heuristics and techniques have been designed towards the goal of achieving convergence to…
Continual Learning (CL) methods usually learn from all available data. However, this is not the case in human cognition which efficiently focuses on key experiences while disregarding the redundant information. Similarly, not all data…
The rich set of interactions between individuals in the society results in complex community structure, capturing highly connected circles of friends, families, or professional cliques in a social network. Thanks to frequent changes in the…
The ability to continuously expand knowledge over time and utilize it to rapidly generalize to new tasks is a key feature of human linguistic intelligence. Existing models that pursue rapid generalization to new tasks (e.g., few-shot…
In democracies, major policy decisions typically require some form of majority or consensus, so elites must secure mass support to govern. Historically, elites could shape support only through limited instruments like schooling and mass…
Large language models has catalyzed the development of personalized dialogue systems, numerous role-playing conversational agents have emerged. While previous research predominantly focused on enhancing the model's capability to follow…
Large Language Model (LLM) agents are increasingly deployed in settings where they interact with a wide variety of people, including users who are unclear, impatient, or reluctant to share information. However, collecting real interaction…
The evolution of large language models (LLMs) toward artificial superhuman intelligence (ASI) hinges on data reproduction, a cyclical process in which models generate, curate and retrain on novel data to refine capabilities. Current…
Altruistic cooperation is costly yet socially desirable. As a result, agents struggle to learn cooperative policies through independent reinforcement learning (RL). Indirect reciprocity, where agents consider their interaction partner's…
Large language models are often ranked according to their level of alignment with human preferences -- a model is better than other models if its outputs are more frequently preferred by humans. One of the popular ways to elicit human…
Complex systems show the capacity to aggregate information and to display coordinated activity. In the case of social systems the interaction of different individuals leads to the emergence of norms, trends in political positions, opinions,…
Personalizing Large Language Models typically relies on static retrieval or one-time adaptation, assuming user preferences remain invariant over time. However, real-world interactions are dynamic, where user interests continuously evolve,…
Recent advancements in large language models (LLMs) have shown remarkable progress, yet their ability to solve complex problems remains limited. In this work, we introduce Cumulative Reasoning (CR), a structured framework that enhances LLM…
Human populations exhibit complex behaviors---characterized by long-range correlations and surges in activity---across a range of social, political, and technological contexts. Yet it remains unclear where these collective behaviors come…
Collective movement is observed widely in nature, where individuals interact locally to produce globally ordered, coherent motion. In typical models of collective motion, each individual takes the average direction of multiple neighbors,…
Cognitive biases are widespread in humans and animals alike, and can sometimes be reinforced by social interactions. One prime bias in judgment and decision-making is the human tendency to underestimate large quantities. Previous research…
Memorization in language models is widely studied but remains difficult to isolate and control. Understanding when and what models memorize is essential for explaining their predictions, yet existing approaches are post-hoc: they can detect…