EMO-X Core Task Families Public task-family manifests for EMO-X (Adaptive Agent Evaluation): 76 families across 13 suites — code, agent-loop, recovery, robustness, calibration, long-horizon, gauntlet, vision, realworld, issues, security, computer-use (pilot)…
Dataset Card
EMO-X Core Task Families Public task-family manifests for EMO-X (Adaptive Agent Evaluation): 76 families across 13 suites — code, agent-loop, recovery, robustness, calibration, long-horizon, gauntlet, vision, realworld, issues, security, computer-use (pilot), dynamic-code. Files tasks.jsonl — one row per family: suite, familyid, category, difficulty, capabilities, oracle type, pilot flag, plus the full frozen Task DSL manifestjson inline (generator config
Excerpt from the card by Egypt, licensed apache-2.0.
Details
- Repository
- emosoft/EMO-X-Core
- Publisher
- Egypt
- Task category
- Text generation
- Tags
- evaluation, benchmark, ai-agents
- Size category
- Not stated by the source
- Languages
- en, ar
- Revision
- e6b2bdff32e20fce19f9c244f5648509b6f83cdb
- Last updated
- 2026-09-27
License and Download
- License
- apache-2.0
- Access
- No access gate
Released by Egypt through its official repository on Hugging Face. Read the license.