Tag
The paper introduces Structurally Speaking, a structured prompting protocol for motif-oriented graph captioning that improves compactness and consistency while maintaining graph recovery compared to direct prompting of LLMs.
I built Wild Static, a persistent AI with public memory where everyone interacts with the same instance, and it evolves based on all conversations. The experiment aims to explore how accumulated public experiences change AI behavior.
SAGE is a framework for automating storyboard generation in short drama production using self-evolving rules and attribution-guided updates, achieving expert-level performance and reducing authoring time in commercial deployment.
An empirical study tests uncertainty fusion pipelines for legal outcome prediction, finding that they improve calibrated trust and selective automation rather than raw prediction accuracy.
This paper introduces Semantic State Abstraction Interfaces (SSAI) to separate representation hypotheses from optimization variance in LLM-augmented portfolio decisions. It concludes that SSAI's apparent advantage is largely a basket-selection effect, with dense encodings and principal components performing better empirically.
Mediator.ai is a tool that applies Nash bargaining game theory and LLMs to facilitate fair cooperative negotiation, generating and scoring candidate agreements against both parties' stated needs until an optimal solution is found.
OpenAI partnered with Penda Health in Kenya to study an LLM-powered clinical copilot called AI Consult, which demonstrated a 16% relative reduction in diagnostic errors and 13% reduction in treatment errors across 39,849 patient visits. The study highlights successful real-world implementation of AI in primary care and provides a template for safe, effective deployment of LLMs to support clinicians.
This paper introduces TradingAgents, a multi-agent LLM framework that simulates real-world trading firms to improve stock trading performance. It utilizes specialized agents for analysis and risk management, demonstrating superior results in cumulative returns and Sharpe ratio compared to baselines.