Optimizing accuracy and performance while eliminating hallucinations of open-domain conversational large language models (LLMs) is an open research challenge. A particularly promising direction is to augment and ground LLMs with information from structured sources. This paper introduces Conversational Tables (cTBLS), a three-step architecture to retrieve and generate dialogue responses grounded on retrieved tabular information. cTBLS uses Transformer encoder embeddings for Dense Table Retrieval and obtains up to 125% relative improvement over the retriever in the previous state-of-the-art system on the HyrbiDialogue dataset. cTBLS then uses a shared process between encoder and decoder models to perform a coarse+fine tabular knowledge (e.g., cell) ranking combined with a GPT-3.5 LLM response generator to yield a 2x relative improvement in ROUGE scores. Finally, human evaluators prefer cTBLs +80% of the time (coherency, fluency) and judge informativeness to be 4x better than the previous state-of-the-art.

翻译：优化开放域对话式大型语言模型（LLMs）的准确性和性能，同时消除其幻觉现象，是一项开放的研究挑战。一个特别有前景的方向是利用结构化信息来增强和约束LLMs。本文提出了对话式表格（cTBLS），这是一种三步架构，用于检索并生成基于检索到的表格信息的对话响应。cTBLS使用Transformer编码器嵌入进行密集表格检索，在HyrbiDialogue数据集上相比先前最先进系统的检索器，其性能相对提升高达125%。随后，cTBLS采用编码器与解码器模型间的共享过程，执行粗粒度加细粒度的表格知识（例如单元格）排序，并结合GPT-3.5 LLM响应生成器，使ROUGE分数相对提升2倍。最后，人类评估者在80%的情况下更偏好cTBLS（在连贯性和流畅性方面），并判定其信息量相比先前最先进系统提升4倍。