Evaluating LLM-based Agents for Multi-Turn Conversations: A Survey

This survey examines evaluation methods for large language model (LLM)-based agents in multi-turn conversational settings. Using a PRISMA-inspired framework, we systematically reviewed nearly 250 scholarly sources, capturing the state of the art from various venues of publication, and establishing a solid foundation for our analysis. Our study offers a structured approach by developing two interrelated taxonomy systems: one that defines \emph{what to evaluate} and another that explains \emph{how to evaluate}. The first taxonomy identifies key components of LLM-based agents for multi-turn conversations and their evaluation dimensions, including task completion, response quality, user experience, memory and context retention, as well as planning and tool integration. These components ensure that the performance of conversational agents is assessed in a holistic and meaningful manner. The second taxonomy system focuses on the evaluation methodologies. It categorizes approaches into annotation-based evaluations, automated metrics, hybrid strategies that combine human assessments with quantitative measures, and self-judging methods utilizing LLMs. This framework not only captures traditional metrics derived from language understanding, such as BLEU and ROUGE scores, but also incorporates advanced techniques that reflect the dynamic, interactive nature of multi-turn dialogues.

Re-Reading Improves Reasoning in Language Models

Reasoning presents a significant and challenging issue for Large Language Models (LLMs). The predominant focus of research has revolved around developing diverse prompting strategies to guide and structure the reasoning processes of LLMs. However, these approaches based on decoder-only causal language models often operate the input question in a single forward pass, potentially missing the rich, back-and-forth interactions inherent in human reasoning. Scant attention has been paid to a critical dimension, i.e., the input question itself embedded within the prompts. In response, we introduce a deceptively simple yet highly effective prompting strategy, termed question "re-reading". Drawing inspiration from human learning and problem-solving, re-reading entails revisiting the question information embedded within input prompts. This approach aligns seamlessly with the cognitive principle of reinforcement, enabling LLMs to extract deeper insights, identify intricate patterns, establish more nuanced connections, and ultimately enhance their reasoning capabilities across various tasks. Experiments conducted on a series of reasoning benchmarks serve to underscore the effectiveness and generality of our method. Moreover, our findings demonstrate that our approach seamlessly integrates with various language models, though-eliciting prompting methods, and ensemble techniques, further underscoring its versatility and compatibility in the realm of LLMs.

TAPEX: Table Pre-training via Learning a Neural SQL Executor

Recent years pre-trained language models hit a success on modeling natural language sentences and (semi-)structured tables. However, existing table pre-training techniques always suffer from low data quality and low pre-training efficiency. In this paper, we show that table pre-training can be realized by learning a neural SQL executor over a synthetic corpus, which is obtained by automatically synthesizing executable SQL queries. By pre-training on the synthetic corpus, our approach TAPEX dramatically improves the performance on downstream tasks, boosting existing language models by at most 19.5%. Meanwhile, TAPEX has remarkably high pre-training efficiency and yields strong results when using a small pre-trained corpus. Experimental results demonstrate that TAPEX outperforms previous table pre-training approaches by a large margin, and our model achieves new state-of-the-art results on four well-known datasets, including improving the WikiSQL denotation accuracy to 89.6% (+4.9%), the WikiTableQuestions denotation accuracy to 57.5% (+4.8%), the SQA denotation accuracy to 74.5% (+3.5%), and the TabFact accuracy to 84.6% (+3.6%). Our work opens the way to reason over structured data by pre-training on synthetic executable programs.