AI Q&A

LongCite is an open source model that trains large language models (LLMs) to generate accurate answers and precise sentence-level quotations in long text question and answer scenarios. The importance of this technology lies in its ability to improve the accuracy and credibility of question answering systems, allowing users to verify the source of the output information. LongCite supports context lengths up to 128K and provides two models: LongCite-glm4-9b and LongCite-llama3.1-8b, which are trained based on GLM-4-9B and Meta-Llama-3.1-8B respectively.

kotaemon is an open source, RAG (Retrieval-Augmented Generation) model-based tool designed to interact with user documents through a chat interface. It supports multiple language model API providers and local language models, and provides a clean and customizable user interface suitable for end users to perform document Q&A and developers to build their own RAG Q&A process.

Fact Finder is an open source intelligent question answering system that uses language models and knowledge graphs to generate natural language answers and provide evidence. The system generates Cypher queries by calling a language model, queries the knowledge graph for answers, and uses another language model call to generate the final natural language answer. Key benefits of Fact Finder include its ability to provide transparency, allowing users to view queries and evidence, and providing intuitive evidence through visual subgraphs.

spRAG is a RAG (Retrieval-Augmented Generation) framework designed for unstructured data and is particularly good at processing complex text queries, such as financial reports, legal documents, and academic papers. Its accuracy is significantly higher than the traditional RAG baseline model on complex open question and answer tasks, such as the FinanceBench benchmark.

Perplexity has launched two online language models: pplx-7b-online and pplx-70b-online. These two models can utilize the latest information on the Internet to form responses, thus being able to accurately answer time-related questions. Responses from pplx online models are more helpful, accurate and timely.