HippoCamp Benchmark for Contextual Agents on Personal File Systems

HippoCamp is a benchmark for evaluating contextual agents in realistic, device-resident personal computing environments. It focuses on multimodal file management over large personal file systems. The dataset was created by MMMem-org and was last updated in April 2026.

Use Cases

• Benchmarking agent performance on multimodal file retrieval tasks using file system metadata and content.
• Evaluating contextual reasoning of agents based on personal computing environment states and user queries.
• Training agents for multimodal file management actions, referencing file paths, types, and content.

Strengths

• Benchmark designed for realistic, device-resident personal computing environments.
• Focuses on multimodal file management, a distinct niche from web or tool-use benchmarks.

Limitations

• Specific dataset size, row count, and column details are unknown.
• The scope is limited to personal computing file systems, which may not generalize to other agent domains.

Provenance

Source

MMMem-org via Hugging Face.

Collection Method

Human-annotated benchmark, as indicated by platform tags.

Freshness

Last updated on 2026-04-02.

Geography

Region 'us' indicated by platform tags.