MemoryAgentBench Model Card
这是可索引项目报告证据页：它保留 MemoryAgentBench Model Card 的源材料入口、机制线索和限制提醒；正文仍需 reader/editor 与 academic public-copy review 后才能当作最终结论引用。
MemoryAgentBench Model Card
| Field | Value |
|---|---|
| Repository | HUST-AI-HYZ/MemoryAgentBench |
| Category | Incremental Agent Memory Benchmark |
| Stars / forks snapshot | 341 / 53 |
| Language | Python |
| License | MIT |
| Raw capture | raw-github/hust-ai-hyz_memoryagentbench.md |
| Updated by | hourly public metadata update, 2026-05-25 |
1. Role in Self Evolve
MemoryAgentBench 是 ICLR 2026 记忆评测代码库，用增量多轮交互测试 agent memory 的准确检索、测试时学习、长程理解和冲突解决能力。
2. Working Principle
incremental multi-turn interaction -> memory injection -> repeated queries -> retrieval/learning/conflict metrics -> agent memory comparison
3. Evidence Path
web GitHub page observed 24 commits, MIT license, Python primary language, ICLR 2026 memory-agent benchmark, LongMemEval/EventQA/FactConsolidation signals, mem0/letta/cognee method folders, 341 stars and 53 forks. Shell GitHub API access remained blocked by DNS and local gh auth was invalid, so this card treats the current snapshot as web-observed rather than API-verified.
4. Teaching Use
Use this card to explain Incremental Agent Memory Benchmark in the raw -> classification -> project card -> site/report pipeline. The reading path is: raw capture -> classification row -> public site card -> project report -> aggregate GitHub analysis.
5. Limits
当前未克隆源码，未运行 benchmark、SDK examples、skill install flows、memory experiments 或 production deployments；star/fork/commit 快照来自公开 GitHub 页面文本或可见页面片段。