Date of Completion

5-13-2026

Degree Type

Honors Thesis

Discipline

Computer Science (CMSI)

First Advisor

Andrew Forney

Abstract

Synthergy is an online social deduction game designed to enable comparative analysis of how large language model-powered agents engage in social deduction and deception under conditions of asymmetric information. Inspired by social deduction games such as Town of Salem, Throne of Lies, and Mafia, the game consists of two factions, Harmony and Discord, to which agents are secretly assigned. Agents must infer others’ affiliations through dialogue, in-game abilities, and voting behavior. To evaluate agent behavior, we conducted 100 simulated games across six agent types: a random baseline agent (RandomSynth), an LLM-based agent (Synth), a chain-of-thought agent (CoT Synth), a Bayesian Belief Network agent (BBN Synth), a Bayesian Belief Network CoT agent (BBN-CoT Synth), and a Causal Decision Network agent (CDN Synth). This paper investigates whether augmenting LLM-powered agents with progressively complex reasoning tools would produce measurable improvements in strategic performance and reasoning coherence within Synthergy. It concludes that advanced reasoning methods, such as the implementation of Bayesian Belief Networks and Causal Decision Network, help enhance what the agents know, but the structured step-by-step prompting enabled by the CoT agents is needed to combine that knowledge with action for successful agent performance. We position this work as a research platform to further the study of deception mechanisms, social deduction, and causal reasoning in multi-agent AI systems.

Download

Share

COinS