AgentsNet: Graph Instances for Multi-Agent Reasoning Benchmark

AgentsNet is a benchmark for multi-agent reasoning, designed to measure collaborative strategy formation, self-organization, and communication within a network topology. The dataset contains graph instances used in the 'AgentsNet: Coordination and Collaborative Reasoning in Multi-Agent LLMs' paper, authored by disco-eth and last updated on 2025-07-17. It draws inspiration from classical problems in distributed systems and graph theory.

Use Cases

• Benchmarking multi-agent LLM performance based on collaborative problem-solving tasks
• Studying self-organization and communication strategies based on network topology constraints
• Developing algorithms for distributed coordination based on graph-theoretic problem instances

Strengths

• Dataset is directly linked to a specific research paper, providing clear academic context.
• Last update timestamp (2025-07-17) is provided, indicating recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.

Provenance

Source

disco-eth

Collection Method

Created for the 'AgentsNet: Coordination and Collaborative Reasoning in Multi-Agent LLMs' research paper.

Freshness

Last updated 2025-07-17 05:53:58