Advances in Large Language Models for Social Simulation and Hallucination Detection

The field of large language models (LLMs) is rapidly advancing, with a focus on improving their ability to simulate social interactions and detect hallucinations. Recent research has demonstrated the potential of LLMs to generate realistic multi-user discussions, simulate online communities, and analyze trust, polarization, and susceptibility to deceptive content in complex social systems. Additionally, there is a growing interest in developing methods to detect and correct hallucinations in LLM-generated text, with a focus on improving factual consistency and reliability. Noteworthy papers in this area include: The Polite Liar: Epistemic Pathology in Language Models, which highlights the issue of confident fabrication in LLMs and proposes an epistemic alignment principle to reward justified confidence over perceived fluency. HalluClean: A Unified Framework to Combat Hallucinations in LLMs, which introduces a lightweight and task-agnostic framework for detecting and correcting hallucinations in LLM-generated text. SynClaimEval: A Framework for Evaluating the Utility of Synthetic Data in Long-Context Claim Verification, which provides a framework for evaluating the utility of synthetic data in long-context claim verification and demonstrates the potential of synthetic data to improve verification and explanation quality.

Sources

Simulating Misinformation Vulnerabilities With Agent Personas

The Polite Liar: Epistemic Pathology in Language Models

Critical Confabulation: Can LLMs Hallucinate for Social Good?

The Collective Turing Test: Large Language Models Can Generate Realistic Multi-User Discussions

What About the Scene with the Hitler Reference? HAUNT: A Framework to Probe LLMs' Self-consistency Via Adversarial Nudge

HalluClean: A Unified Framework to Combat Hallucinations in LLMs

Sure! Here's a short and concise title for your paper: "Contamination in Generated Text Detection Benchmarks"

SynClaimEval: A Framework for Evaluating the Utility of Synthetic Data in Long-Context Claim Verification

Built with on top of