Skip to content

Dataset · ai-hyz

MemoryAgentBench for AI Agent Memory Evaluation

ai-hyz/MemoryAgentBench

An English dataset (up to 1K examples) for testing agent long-term memory.

What's inside

A compact English text dataset with a volume of less than 1,000 examples. It contains question-answer pairs (including a subset from DetectiveQA and tasks from Longmemeval), as well as key points for testing context retention. The dataset is regularly updated by the authors to remove ineffective samples.

What you can do with it

  • Evaluating long-term memory and long-context handling in AI agents
  • Testing models on QA and classification tasks relying on history
Tasks
Question answering, Zero shot classification, Summarization, Text classification, Text generation
Languages
English
Size
up to 1K rows
Downloads this month
11.1K
Likes
44

How to load it

from datasets import load_dataset

ds = load_dataset("ai-hyz/MemoryAgentBench")