![](https://private-user-images.githubusercontent.com/43441551/248049973-2aa43d6b-ae4a-4fc3-9f23-b848a530c5d4.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.EPJ-t1XK9eEPEtLiW6V-5W2sW2Il_iEUbQAq0ZwWPi4)
Multi-Skill Advancer(MSA) is a benchmark designed to ensure large language models (LLMs) simultaneously retain specialized knowledge and enhance general performance. It embodies the concept of cultivating a versatile hand, promoting broader understanding without forgetting area of expertise.