Kirkland, Washington – Appen Limited (ASX:APX), a prominent provider of high-quality data for the AI lifecycle, has unveiled two new products that empower users to deploy large language models (LLMs) with exceptional performance. These models are designed to produce responses that are helpful, harmless, and unbiased, thus reducing bias and toxicity. The products, AI Chat Feedback and Benchmarking, address the growing demand for advanced conversational AI and aid in contextual understanding, coherence, and model selection. The rise of LLM-powered chatbots and assistants has spurred the need for more sophisticated conversational AI capable of handling various tasks. Testing the contextual understanding and coherence of LLMs in complex conversations that extend over multiple turns or dialogues is crucial for real-world applications. This assessment helps identify strengths and weaknesses in managing extended interactions, resulting in improved user experiences and practical utility of the model. Appen's AI Chat Feedback streamlines the data flow for multiple rounds of evaluation, providing customers with the necessary data to enhance their models. Meanwhile, Appen's Benchmarking tool assists businesses in selecting the most suitable LLM for their specific enterprise application, considering factors such as user experience, ease of maintenance, and profitability. This solution allows customers to evaluate various models based on commonly used or fully customized dimensions. Moreover, with the help of Appen's AI Training Specialists, performance evaluation can be conducted based on demographic dimensions of interest, including gender, ethnicity, and language. A configurable dashboard enables efficient comparison of multiple models across different dimensions. Appen CEO Armughan Ahmad emphasized the importance of releasing AI chatbots that are trustworthy and not harmful to the public. He stated that Appen's new evaluation products provide a crucial trust layer, backed by robust datasets, proven processes, and a team of over a million human experts with years of AI training experience. The performance of LLM models has been shown to greatly benefit from human feedback.

Appen's technology is augmented by a global crowd of more than 1 million AI Training Specialists responsible for evaluating datasets for accuracy and bias. The AI Chat Feedback tool facilitates direct communication between specialists and LLMs, enabling learning from diverse, natural chat data. Appen's expertise in intuitive annotation platforms has resulted in a chat interface that promotes familiarity and ease of use. Specialists engage in live conversations with models, rating, flagging, and providing context for evaluation. A dedicated project staff meticulously analyzes each batch of data, ensuring optimal data quality by identifying edge cases and making necessary optimizations. Appen is continuously improving its products to provide more advanced AI capabilities in the future. To learn more about Appen's new products, please visit their website at Appen. com or get in touch with their sales team. Appen is a global leader in providing data for the AI lifecycle, with extensive experience in data sourcing, annotation, and model evaluation. With their expertise, platform, and global crowd, Appen empowers organizations to launch groundbreaking AI products quickly and at scale. Appen's highly advanced AI-assisted data annotation platform, along with its diverse global crowd of more than 1 million contributors who speak over 235 languages, make them a trusted partner for technology, automotive, finance, retail, healthcare, and government sectors. Appen serves clients worldwide and is committed to offering the most secure and exceptional website experience by recommending the latest versions of Chrome, Firefox, and Safari over Internet Explorer, which will no longer be supported after August 17, 2021.