Large Language Models are rapidly evolving from text generators into reasoning systems that can act as autonomous agents. When placed in social contexts, these agents display emergent behaviors such as forming coalitions, spreading information, and making collective decisions.
Generative AI for Pro-Democracy Platforms · From Novel Chemicals to Opera
Online discourse faces challenges in facilitating substantive and productive political conversations. Recent technologies have explored the potential of generative AI to promote civil discourse, encourage the development of mutual understanding in a discussion, produce . . .
The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
Recent generations of frontier language models have introduced Large Reasoning Models (LRMs) that generate detailed thinking processes…
KARMA: Leveraging Multi-Agent LLMs for Automated Knowledge Graph Enrichment
Maintaining comprehensive and up-to-date knowledge graphs (KGs) is critical for modern AI systems, but manual curation struggles to scale with the rapid growth of scientific literature. This paper presents KARMA, a novel framework employing multi-agent large language models (LLMs) to automate KG enrichment through structured analysis of unstructured text. Our approach employs nine collaborative agents, spanning entity discovery, relation extraction, schema alignment, and conflict resolution that iteratively parse documents, verify extracted knowledge, and integrate it into existing graph structures while adhering to domain-specific schema. Experiments on 1,200 PubMed articles from three different domains demonstrate the effectiveness of KARMA in knowledge graph enrichment, with the identification of up to 38,230 new entities while achieving 83.1% LLM-verified correctness and reducing conflict edges by 18.6% through multi-layer assessments.
LLM-Augmented Agent-Based Simulation of Cyber Social Agents
Social media bots are often portrayed as malicious automations to be detected and removed, but automation on social media platforms encompasses a wide spectrum of roles, behaviors, and rhetorical strategies that can shape online discourse in both constructive and destructive ways. This work conceptualizes these automated actors as \textbf{Cyber Social Agents} through a taxonomy of archetypes characterized by their operational tactics and rhetorical style. We integrate this taxonomy into an LLM-driven social simulation, where agents communicate and coordinate in a network environment. Using empirically grounded CSA personas, we simulate heterogeneous online societies to study how agent diversity shapes information diffusion and social influence. Our approach bridges the gap between rule-based agent models and LLM-powered simulations with generated agents that are parameterized from empirical data.
Stanford CS336 Language Modeling from Scratch I 2025
If LLMs Have Human-Like Attributes, Then So Does Age of Empires II
Much research has been carried out on large language models (LLMs) and LLM-powered agentic workflows. However, many works within the field state emergence of, ascribe to, or assume, generalised anthropomorphic attributes to them (e.g., morality or understanding of natural language). Our goal is not to argue in favour or against the existence of these attributes, but to point out that these conclusions could be incorrect. For this we build and train a simple neural network on the videogame Age of Empires II, and note that any entity in a sufficiently-powerful substrate, such as LEGO or the Greater Boston Area, could also present such attributes. Hence, the purported anthropomorphic attributes of LLMs are empirically non-unique: although some properties (e.g., responses to prompts) could remain constant, others, such as the interpretation of their perceived behaviour, might change with the substrate. Thus, any empirically-grounded discussion requires explicit measurement criteria; otherwise the interpretation is left to the representation. We then show that assuming that these attributes exist or not in a system, independent of the substrate and in a generalised way, leads to either circular or uninformative conclusions, regardless of the experimenter's viewpoint on the subject. Finally we propose a 'null' assumption, where one assumes LLM non-uniqueness instead of assuming anthropomorphic attributes to set up an experiment, along with examples of it. We also discuss potential objections to our work, briefly survey the field, and prove that Age of Empires II is functionally- and Turing-complete.
Language models are multiverse generators
"Actualities seem to float in a wider sea of possibilities from out of which they were chosen; and somewhere, indeterminism says, such possibilities exist, and form part of the truth."
IO Factory: Simulating AI-Enabled Influence Campaigns at Scale
We introduce IO Factory, an AI-driven framework for simulating information and influence campaigns as fully integrated, traceable processes. The threat of digital manipulation now extends beyond persuasive text from individual language models to AI swarms, i.e., persistent groups of coordinated agents that adapt to platform feedback and disguise organized campaigns as ordinary social interaction. Because such campaigns cannot be identified from isolated messages alone, they must be analyzed across a continuous spectrum of planning, platform action, exposure, interpretation, measurement, and adaptation. IO Factory represents this process inside a controlled simulated platform, linking actor roles, platform actions, exposure records, structured model-based evaluations, and configured changes in the simulated population. We implement the architecture and evaluate it across configurations of up to 100,000 agents. The results show that IO Factory executes campaign timelines at scale and produces inspectable evidence of exposure and measured movement in configured belief variables. By recording the actors, objectives, action constraints, exposure paths, and measurement rules used in each run, IO Factory supports reproducible research and red-team analysis of coordinated influence.
We Don't Understand Neural Networks At The Algorithmic Level
The largest ongoing debate about AI is “Are Large Language Models (LLMs) intelligent?” That makes sense, at least: the evidence is ambiguous and the stakes a...
Bridging Systems: Open Problems for Countering Destructive Divisiveness across Ranking, Recommenders, and Governance
Divisiveness appears to be increasing in much of the world, leading to concern about political violence and a decreasing capacity to collaboratively address large-scale societal challenges. In this working paper we aim to articulate an interdisciplinary research and practice area focused on what we call bridging systems: systems which increase mutual understanding and trust across divides, creating space for productive conflict, deliberation, or cooperation. We give examples of bridging systems across three domains: recommender systems on social media, collective response systems, and human-facilitated group deliberation. We argue that these examples can be more meaningfully understood as processes for attention-allocation (as opposed to "content distribution" or "amplification") and develop a corresponding framework to explore similarities - and opportunities for bridging - across these seemingly disparate domains. We focus particularly on the potential of bridging-based ranking to bring the benefits of offline bridging into spaces which are already governed by algorithms. Throughout, we suggest research directions that could improve our capacity to incorporate bridging into a world increasingly mediated by algorithms and artificial intelligence.
The assistant axis: situating and stabilizing the character of large language models
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
AutoSchemaKG: Autonomous Knowledge Graph Construction through Dynamic Schema Induction from Web-Scale Corpora
In the current era of information abundance, transforming vast amounts of unstructured data into structured, machine-readable knowledge remains one of the most significant challenges in artificial intelligence. Knowledge Graphs (KGs) have emerged as the cornerstone technology for this transformation Zhao et al. (2024), providing the semantic backbone for applications ranging from search engines and question answering Wu et al. (2024); Chen et al. (2024c); Zong et al. (2024); Sun et al. (2024b) to recommendation systems Lyu et al. (2024) and complex reasoning tasks Li et al. (2024b). Yet despite their critical importance, current KG construction approaches remain hampered by an inherent paradox: they require predefined schemas created by domain experts, which fundamentally limits their scalability, adaptability, and domain coverage.
Scaling Laws Across Model Architectures: A Comparative Analysis of...
The scaling of large language models (LLMs) is a critical research area for the efficiency and effectiveness of model training and deployment. Our work investigates the transferability and...
Wikipedia:Signs of AI writing
This is a list of writing and formatting conventions typical of AI chatbots such as ChatGPT, with real examples taken from Wikipedia articles and drafts. It is a field guide to help detect undisclosed AI-generated content on Wikipedia: while some of the signs may be broadly applicable, some may not apply in a non-Wikipedia context.[a] Not all text featuring these indicators is AI-generated, as the large language models that power AI chatbots are trained on human writing, including Wikipedia. Many elements of AI writing can be found in editorials, blogs, or fan fiction.
Self-Instruct: Aligning Language Models with Self-Generated Instructions
Large "instruction-tuned" language models (i.e., finetuned to respond to instructions) have demonstrated a remarkable ability to generalize zero-shot to new tasks. Nevertheless, they depend...