Meta 将基于 Muse Wonder 训练的专用模型上线 Muse 与 Meta AI
Bringing Specialized Models to Muse and Meta AI
Meta 宣布把基于 Muse Wonder 训练的专用模型推向 Muse 和 Meta AI 用户，Personal Superintelligence Index 得分从 38% 提升到 47%。
原文给出了 Active Memory 与 Active Focus 的具体评测数字，读者可据此评估 Muse 与 Meta AI 在记忆和任务坚持上的实际变化。
A few weeks ago, we introduced the Muse Wonder series, our most capable models to date, delivering state-of-the-art performance on personal and professional agentic tasks. Today, we’re introducing specialized versions of our existing models, trained using Muse Wonder, and rolling them out to people who use Muse and Meta AI.
Bringing Key Frontier Capabilities to Muse and Meta AI
We designed Muse Wonder to have frontier capabilities for complex tasks at work and in everyday life. We’ve also used it to improve the models behind Muse and Meta AI, bringing some of those capabilities to even more people.
Personal Superintelligence Index
Current and previous Muse models on the Personal Superintelligence Index: 47% and 38%.. Index Score
| Series | Measurement | Value |
|---|---|---|
| Current Muse Model | Model | Current Muse Model |
| Current Muse Model | Index Score | 47% |
| Previous Muse Model | Model | Previous Muse Model |
| Previous Muse Model | Index Score | 38% |
Remembering What Matters
With Muse Wonder, we introduced Active Memory and Active Focus. We’re bringing both of these to our specialized models in Muse and Meta AI.
Active Memory enables Muse and Meta AI to remember what matters and recognize when earlier information becomes relevant, making it easier to build off your past conversations. We’ve trained our models to recognize what type of information is temporary versus permanent, to reason across time, and to distinguish between what is still relevant and what has become stale. For example, our improved models in Muse show a 2.4x reduction in forgetting critical information necessary for helping you achieve your goals.
Active Memory Eval
Current and previous Muse models on the Active Memory eval: 90% and 76%.. Tasks Correctly Using Critical Context
| Series | Measurement | Value |
|---|---|---|
| Current Muse Model | Model | Current Muse Model |
| Current Muse Model | Tasks Correctly Using Critical Context | 90% |
| Previous Muse Model | Model | Previous Muse Model |
| Previous Muse Model | Tasks Correctly Using Critical Context | 76% |
Working Toward Your Goals
Active Focus enables Muse to continue working toward a goal, even when faced with unexpected challenges. We’ve trained our models on a variety of bottlenecks reflective of the real world, so that they can try harder, recover, or find alternatives on your behalf. In Muse, our models show a 67% improvement in completing everyday tasks where unexpected issues arise.
We’ve also developed our specialized models to leverage Muse Wonder’s performance on behaviors that are especially important to Muse. For example, they learned to be excellent multi-taskers that know when to push on your goals independently or collaborate live with you. They can also better adapt to steers and clarifications in the middle of running tasks without getting confused.
In Meta AI, the same improvements make for a better thinking partner and collaborator. It can break complex problems into clear steps, ask sharper follow-up questions, and keep up when you change direction mid-conversation.
By adding these specialized models to Muse and Meta AI, we aim to give more people access to even more capable superintelligence. We can’t wait for you to experience it for yourself.
来源：Meta AI：Research Blog · research.meta.ai