LongCite is an open-source model that trains large language models (LLMs) to produce accurate responses and precise sentence-level citations in long text question-answering scenarios. Its significance lies in enhancing the accuracy and credibility of question-answering systems, allowing users to verify the sources of the output information. LongCite supports context lengths of up to 128K and provides two models: LongCite-glm4-9b and LongCite-llama3.1-8b, trained respectively on GLM-4-9B and Meta-Llama-3.1-8B.

Fact Finder is an open-source intelligent question-answering system that utilizes language models and knowledge graphs to generate natural language answers and provide evidence. The system generates Cypher queries by invoking language models, queries the knowledge graph for answers, and uses another language model to produce the final natural language response. The main advantages of Fact Finder include the ability to provide transparency, allowing users to see both queries and evidence, as well as offering intuitive evidence through visual subgraphs.

ChatQA is a series of conversational question answering (QA) models that can achieve GPT-4 level accuracy. We propose a two-stage instruction tuning method that can significantly improve the zero-shot conversational QA results of large language models (LLMs). To handle retrieval in dialogue-based QA, we fine-tuned dense retrievers on multi-turn QA datasets, which provides comparable results to using state-of-the-art query rewrite models while significantly reducing deployment costs. Notably, our ChatQA-70B outperforms GPT-4 (54.14 vs. 53.90) on the average score across 10 conversational QA datasets without relying on any synthetic data from the OpenAI GPT model.

LongCite is an open-source model that trains large language models (LLMs) to produce accurate responses and precise sentence-level citations in long text question-answering scenarios. Its significance lies in enhancing the accuracy and credibility of question-answering systems, allowing users to verify the sources of the output information. LongCite supports context lengths of up to 128K and provides two models: LongCite-glm4-9b and LongCite-llama3.1-8b, trained respectively on GLM-4-9B and Meta-Llama-3.1-8B.

Fact Finder is an open-source intelligent question-answering system that utilizes language models and knowledge graphs to generate natural language answers and provide evidence. The system generates Cypher queries by invoking language models, queries the knowledge graph for answers, and uses another language model to produce the final natural language response. The main advantages of Fact Finder include the ability to provide transparency, allowing users to see both queries and evidence, as well as offering intuitive evidence through visual subgraphs.

ChatQA is a series of conversational question answering (QA) models that can achieve GPT-4 level accuracy. We propose a two-stage instruction tuning method that can significantly improve the zero-shot conversational QA results of large language models (LLMs). To handle retrieval in dialogue-based QA, we fine-tuned dense retrievers on multi-turn QA datasets, which provides comparable results to using state-of-the-art query rewrite models while significantly reducing deployment costs. Notably, our ChatQA-70B outperforms GPT-4 (54.14 vs. 53.90) on the average score across 10 conversational QA datasets without relying on any synthetic data from the OpenAI GPT model.