NEWSLETTER
A recent advance in the field of Artificial Intelligence is the introduction of Large Language Models (LLM). These models allow us to understand language more concisely and thus make the best use of Natural Language Processing (NLP) and Natural Language Understanding (NLU). These models work well for all other tasks including summarizing text, answering questions, generating content, language translation, etc. They understand complex text prompts, including texts with reasoning and logic, and identify patterns and relationships among those data.
Although language models have shown incredible performance and have developed significantly in recent times by proving their proficiency in a variety of tasks, they still find it difficult to use tools via API calls efficiently. Even famous LLMs like GPT-4 struggle to generate accurate input arguments and often recommend inappropriate API calls. To address this issue, the Berkeley and Microsoft Research researchers proposed Gorilla, a refined LLaMA-based model that outperforms GPT-4 in terms of API call throughput. Gorilla helps choose the right API, improving the ability of LLMs to work with external tools to carry out particular activities.
The research team also created an APIBench dataset, which is made up of a sizeable corpus of APIs with overlapping functionality. The dataset was created by collecting public model hubs such as TorchHub, TensorHub, and HuggingFace for their ML APIs. Every TorchHub and TensorHub API request is listed for each API, and the top 20 HuggingFace models are chosen for each task category. In addition, they produce ten dummy user query requests for each API using the self-instruction method.
Using this APIBench data set and document retrieval, the researchers have perfected Gorilla. Gorilla, the 7 billion parameter model outperforms GPT-4 in correctness of API performance and reduces hallucinatory errors. The effective integration of the document retriever with Gorilla demonstrates the potential for LLMs to use the tools with greater precision. Gorilla’s enhanced API call generation capabilities and ability to modify documentation as needed improves the applicability and reliability of model results. This development is important because it allows LLMs to keep up with regularly updated documentation, providing users with more accurate and up-to-date information.
One of the examples shared by the researchers shows how Gorilla correctly recognizes tasks and delivers fully qualified API results. API calls generated by the models showed that GPT-4 produced API requests for what-if models, demonstrating a lack of understanding of the task. Claude chose the wrong library, showing a lack of ability to recognize the correct resources. Gorilla, on the other hand, correctly recognized the task. Thus, Gorilla differs from GPT-4 and Claude in that API call creation is precise, demonstrating both its improved performance and task understanding.
In conclusion, Gorilla is an important addition to the list of language models, as it even addresses the problem of writing API calls. Its capabilities allow for the reduction of hallucination-related issues and reliability.
review the Paper, GitHub linkand Project page. Don’t forget to join our 22k+ ML SubReddit, discord channel, and electronic newsletter, where we share the latest AI research news, exciting AI projects, and more. If you have any questions about the article above or if we missed anything, feel free to email us at asif@marktechpost.com
Check out 100 AI tools at AI Tools Club
Tanya Malhotra is a final year student at the University of Petroleum and Power Studies, Dehradun, studying BTech in Computer Engineering with a specialization in Artificial Intelligence and Machine Learning.
She is a data science enthusiast with good analytical and critical thinking, along with a keen interest in acquiring new skills, leading groups, and managing work in an organized manner.
Definitive Guide to Data Labeling in Machine Learning 
