GateMem: A Benchmark for Evaluating Memory Governance in Multi-Agent Systems

GateMem is a benchmark for evaluating memory governance in multi-principal shared-memory agents. It assesses three dimensions: Utility, Access Control, and Active Forgetting. The dataset was created by Ray368 and last updated on 2026-06-22.

Use Cases

• Benchmarking agent utility based on answering authorized requests
• Evaluating access control mechanisms based on preventing unauthorized information disclosure
• Testing active forgetting capabilities based on avoiding recovery of protected information
• Comparing multi-agent system architectures based on memory governance performance

Strengths

• Focuses on three specific evaluation dimensions: Utility, Access Control, and Active Forgetting
• Last updated on 2026-06-22, indicating recent maintenance

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment
• Description metadata is limited; actual data quality requires manual inspection after download

Provenance

Source

huggingface

Freshness

Last updated 2026-06-22 04:36:59