datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Terminal-Bench-Hard
Terminal-Bench Hard
Terminal-Bench Hard is a set of 100 challenging terminal-based agent tasks.
The tasks cover software engineering, debugging, data processing, system
administration, security, scientific computing, and related command-line
workflows.
Contents
tasks/: runnable tasks in Harbor format.
metadata/tasks.parquet: searchable task metadata and instructions.
Each task directory contains task.toml, instruction.md, an
environment/ directory, and verifier… See the full description on the dataset page: https://huggingface.co/datasets/Zhongzhi1228/Terminal-Bench-Hard.ValidMem
ValidMem
ValidMem is a 466-case diagnostic benchmark for validity-aware agent
memory: whether a memory system can tell that a stored fact is no longer
true, and still surface it when the user asks about the past.
It accompanies the EMNLP 2026 paper Validity Is Not Visibility: A Lifecycle
FSM for Agent Memory.
What makes it different from existing memory benchmarks: each fact is
annotated along both invalidation dimensions, and each store is queried
under both intents.… See the full description on the dataset page: https://huggingface.co/datasets/Zhou11Alex/ValidMem.