![[The AI Show Episode 144]: ChatGPT’s New Memory, Shopify CEO’s Leaked “AI First” Memo, Google Cloud Next Releases, o3 and o4-mini Coming Soon & Llama 4’s Rocky Launch](https://www.marketingaiinstitute.com/hubfs/ep%20144%20cover.png)
![Blue Archive tier list [April 2025]](https://media.pocketgamer.com/artwork/na-33404-1636469504/blue-archive-screenshot-2.jpg?#)
.png?#)
.png?width=1920&height=1920&fit=bounds&quality=70&format=jpg&auto=webp#)
.webp?#)
![PSA: It’s not just you, Spotify is down [U: Fixed]](https://i0.wp.com/9to5mac.com/wp-content/uploads/sites/6/2023/06/spotify-logo-2.jpg?resize=1200%2C628&quality=82&strip=all&ssl=1){kind=logo}
![Pixel 9a vs. Pixel 7a: Is now the time to upgrade? [Video]](https://i0.wp.com/9to5google.com/wp-content/uploads/sites/4/2025/04/Pixel-7a-8a-and-9a-design.jpg?resize=1200%2C628&quality=82&strip=all&ssl=1){kind=tracking}
![[Update: Optional] Google rolling out auto-restart security feature to Android](https://i0.wp.com/9to5google.com/wp-content/uploads/sites/4/2025/01/google-play-services-2.jpg?resize=1200%2C628&quality=82&strip=all&ssl=1)
![Apple Releases iOS 18.4.1 and iPadOS 18.4.1 [Download]](https://www.iclarified.com/images/news/97043/97043/97043-640.jpg)
![Apple Releases visionOS 2.4.1 for Vision Pro [Download]](https://www.iclarified.com/images/news/97046/97046/97046-640.jpg)
![Apple Vision 'Air' Headset May Feature Titanium and iPhone 5-Era Black Finish [Rumor]](https://www.iclarified.com/images/news/97040/97040/97040-640.jpg)
![[Update: Trump Backtracks]The U.S. Just Defunded a Key Security Database, And Your Android Phone Could Pay the Price](https://www.androidheadlines.com/wp-content/uploads/2025/03/Android-logo-AM-AH.jpg){kind=logo}
Legal AI Beats GPT-4o on Bar Exam: LegalLlama-8B Achieves 70% Accuracy
This is a Plain English Papers summary of a research paper called Legal AI Beats GPT-4o on Bar Exam: LegalLlama-8B Achieves 70% Accuracy. If you like these kinds of analysis, you should join AImodels.fyi or follow us on Twitter. Overview Llama 3.1 model fine-tuned specifically for legal reasoning Achieved 69.9% accuracy on Multi-state Bar Exam (MBE) questions Outperformed GPT-4o (66.6%) with much smaller model size Used efficient supervised fine-tuning (SFT) with only 2,700 legal questions Demonstrated cost-effective approach to specialized domain adaptation Created LegalLlama-8B model with minimal legal expertise required Plain English Explanation Law school graduates must pass the bar exam to practice law. This challenging test includes the Multi-state Bar Exam (MBE), which tests legal reasoning through multiple-choice questions. Researchers created a specialized AI system called [LegalLlama-8B](https://aimodels.fyi/p... Click here to read the full summary of this paper
This is a Plain English Papers summary of a research paper called Legal AI Beats GPT-4o on Bar Exam: LegalLlama-8B Achieves 70% Accuracy. If you like these kinds of analysis, you should join AImodels.fyi or follow us on Twitter.
Overview
- Llama 3.1 model fine-tuned specifically for legal reasoning
- Achieved 69.9% accuracy on Multi-state Bar Exam (MBE) questions
- Outperformed GPT-4o (66.6%) with much smaller model size
- Used efficient supervised fine-tuning (SFT) with only 2,700 legal questions
- Demonstrated cost-effective approach to specialized domain adaptation
- Created LegalLlama-8B model with minimal legal expertise required
Plain English Explanation
Law school graduates must pass the bar exam to practice law. This challenging test includes the Multi-state Bar Exam (MBE), which tests legal reasoning through multiple-choice questions.
Researchers created a specialized AI system called [LegalLlama-8B](https://aimodels.fyi/p...
