Large Language Models (LLMs) have successfully catered their way into the difficult areas of Artificial Intelligence. With their amazing ability to provide unique and artistic content with great linguistic accuracy and consistency, LLMs are helping out in every industry. Large Language Models are sometimes augmented with reasoning skills and the power to make use of different tools. Augmentation principally refers to enhancing or expanding by adding additional elements or features. Augmented LLMs are those which are added with external tools and skills with a purpose to increase their performance in order that they perform beyond their inherent capabilities.

Applications like Auto-GPT for autonomous task execution have been made possible by Augmented Language Models (ALMs) only. Current ALM attempts mostly depend on the prompting paradigm with interleaved verbal reasoning and tool-calling, which have been effective but in addition imposes certain limitations. When connecting with external tools, it first necessitates the regular execution and suspension of LLMs, which causes delays and increases token usage. Secondly, LLMs generate tokens based on the previous context, and when halted for tool response, they resume token generation by feeding all historical tokens, which ends up in significant prompt redundancy resulting in high cost when it comes to token consumption for business LLM services.

To handle the challenges, recently, a team of researchers has proposed ReWOO (Reasoning WithOut Statement), a modular paradigm to scale back token consumption. The thought behind ReWOO is to separate the reasoning means of the LLM from external observations, which might help reduce the token consumption significantly. ReWOO minimizes the computational load related to repeated prompts by separating the reasoning process from external observations.

The important thing components of an ALM are step-wise reasoning, tool calls, and summarization, which ReWOO divides into three separate modules: Planner, Employee, and Solver. The Planner breaks down a task and formulates a blueprint of interdependent plans, that are each assigned to a Employee. The Employee retrieves external knowledge from tools to offer evidence, and the Solver synthesizes all of the plans and evidence to provide the ultimate answer to the initial task to be accomplished.

To guage ReWOO’s performance, the team has carried out an intensive evaluation across six open Natural Language Processing (NLP) benchmarks and a curated dataset. The outcomes consistently showed improvements with the proposed methodology, with ReWOO achieving a 5× token efficiency gain and a 4% accuracy improvement on the HotpotQA benchmark, which involves multi-step reasoning tasks. ReWOO also proved to be robust in situations where the external tools had failure issues.

The decoupling of parametric modules from nonparametric tool calls not only increases prompt efficiency but in addition enables instruction fine-tuning in ReWOO. A 175B parameter GPT3.5 can have its reasoning capability offloaded to a smaller language model, 7B LLaMA, through fine-tuning, resulting in a major reduction in model parameters, which highlights the potential for developing effective and scalable ALMs.

Consequently, ReWOO is a promising modular paradigm for ALMs as, for the primary time, it overcomes the challenges of redundant prompts and computation complexity.

Check Out The Paper and Github link. Don’t forget to hitch our 22k+ ML SubReddit, Discord Channel, and Email Newsletter, where we share the newest AI research news, cool AI projects, and more. If you have got any questions regarding the above article or if we missed anything, be at liberty to email us at Asif@marktechpost.com

🚀 Check Out 100’s AI Tools in AI Tools Club