EDGE: Engine for Deterministic Graph Evaluation through Conversation Simulation from Graph Structured DSL Configuration
This paper introduces a formal evaluation methodology that is grounded in AgentGraph, a planner powered by a domain specific language that represents agent reasoning through a dynamically adjustable directed graph, and defines novel metrics that measure response and trajectory determinism, structural adherence and semantic consistency across both exact replays and their linguistic variants.
Ram Kulathumani, Regunathan Radhakrishnan, Anupam Tripathi et al.
· 0 citations