Large language models (LLMs) are increasingly used for complex multi-turn conversations across diverse real-world applications. However, existing benchmarks mainly focus on single-turn evaluations, overlooking the models’ capabilities in multi-turn interactions. To address this gap, we introduce , a comprehensive benchmark to evaluate the multi-turn conversational abilities of LLMs. By analyzing human-LLM conversations, we categorize interaction patterns into four types: recollection, expansion, refinement, and follow-up. We construct multi-turn queries for each category either by augmenting existing datasets or creating new examples using GPT-4 with a human-in-the-loop process to avoid data leakage. To study the factors impacting multi-turn abilities, we create single-turn versions of the 1170 multi-turn queries and compare performance. Our evaluation of 10 well-known LLMs shows that while closed-source models generally surpass open-source ones, certain open-source models exceed GPT-3.5-Turbo in specific tasks. We observe significant performance degradation in multi-turn settings compared to single-turn settings in most models, which is not correlated with the models’ fundamental capabilities. Moreover, we identify the distance to relevant content and susceptibility to error propagation as the key factors influencing multi-turn performance.

Managing long sequences has become an important and necessary feature for large language models (LLMs). However, assessing their ability to handle long contexts remains a challenge. This paper introduces M⁴LE, a Multi-ability, Multi-range, Multi-task, Multi-domain benchmark for Long-context Evaluation. It encompasses 36 NLP datasets, covering 11 types of tasks and 12 domains, providing a comprehensive test bed. To address the lack of tasks featuring naturally long sequences, we propose an automatic approach to convert short-sequence tasks into long-sequence scenarios. These scenarios evaluate LLMs’ long-context understanding across five key abilities: understanding of single or multiple relevant spans in long contexts based on explicit or semantic hints, and global context understanding. This automatic approach allows us to create instances evenly distributed from 1k to 8k input length. Our evaluation of 11 prominent LLMs reveals that 1) Current LLMs struggle to understand long context, particularly when tasks require multiple-span attention. 2) Semantic retrieval is more difficult for competent LLMs. 3) Models fine-tuned on longer text with position interpolation have comparable performance to those using Neural Tangent Kernel (NTK) aware scaling methods without fine-tuning. We make our benchmark publicly available to encourage future research in this challenging area.

Dialogue policy learning (DPL) aims to determine an abstract representation (also known as action) to guide what the response should be. Typically, DPL is cast as a sequential decision problem across a series of predefined action candidates. However, such static and narrow actions can limit response diversity and impede the dialogue agent’s adaptability to new scenarios and edge cases. To overcome these challenges, we introduce a novel Joint Transformer Reinforcement Learning framework, coined as JoTR, where a text-to-text Transformer-based model is employed to directly generate dialogue actions. More concretely, JoTR formulates a token-grained policy, facilitating more dynamic and adaptable dialogue action generation without the need for predefined action candidates. This method not only enhances the diversity of responses but also significantly improves the system’s capability to manage unfamiliar scenarios. Furthermore, JoTR utilizes Reinforcement Learning with a reward-shaping mechanism to efficiently fine-tune the token-grained policy. This allows the model to evolve through interactions, thereby enhancing its performance over time. Our extensive evaluation demonstrates that JoTR surpasses previous state-of-the-art models, showing improvements of 9% and 13% in success rate, and 34% and 37% in the diversity of dialogue actions across two benchmark dialogue modeling tasks respectively. These results have been validated by both user simulators and human evaluators. Code and data are available at ://github.com/KwanWaiChung/JoTR.

Automatic diagnosis (AD), a critical application of AI in healthcare, employs machine learning techniques to assist doctors in gathering patient symptom information for precise disease diagnosis. The Transformer-based method utilizes an input symptom sequence, predicts itself through auto-regression, and employs the hidden state of the final symptom to determine the disease. Despite its simplicity and superior performance demonstrated, a decline in disease diagnosis accuracy is observed caused by 1) a mismatch between symptoms observed during training and generation, and 2) the effect of different symptom orders on disease prediction. To address the above obstacles, we introduce the CoAD, a novel disease and symptom collaborative generation framework, which incorporates several key innovations to improve AD: 1) aligning sentence-level disease labels with multiple possible symptom inquiry steps to bridge the gap between training and generation; 2) expanding symptom labels for each sub-sequence of symptoms to enhance annotation and eliminate the effect of symptom order; 3) developing a repeated symptom input schema to effectively and efficiently learn the expanded disease and symptom labels. We evaluate the CoAD framework using four datasets, including three public and one private, and demonstrate that it achieves an average 2.3% improvement over previous state-of-the-art results in automatic disease diagnosis. For reproducibility, we release the code and data at https://github.com/KwanWaiChung/coad.

Generating persona consistent dialogue response is important for developing an intelligent conversational agent. Recent works typically fine-tune large-scale pre-trained models on this task by concatenating persona texts and dialogue history as a single input sequence to generate the target response. While simple and effective, our analysis shows that this popular practice is seriously affected by order sensitivity where different input orders of persona sentences significantly impact the quality and consistency of generated response, resulting in severe performance fluctuations (i.e., 29.4% on GPT2 and 83.2% on BART). To mitigate the order sensitivity problem, we propose a model-agnostic framework, ORder Insensitive Generation (ORIG), which enables dialogue models to learn robust representation under different persona orders and improve the consistency of response generation. Experiments on the Persona-Chat dataset justify the effectiveness and superiority of our method with two dominant pre-trained models (GPT2 and BART).

Knowledge probing is a task to assess the knowledge encoded within pre-trained language models (PLMs) by having the PLM complete prompts such as “Italy is located in __,”. The model’s prediction precision serves as a lower bound for the amount of knowledge it contains. Subsequent works explore training a series of vectors as prompts to guide PLMs towards more accurate predictions. However, these methods compromise the readability of the prompts. We cannot directly understand these prompts from their literal meaning, making it difficult to verify whether they are correct. Consequently, the credibility of probing results derived from these prompts is diminished. To address the issue, we propose a novel method called ReadPrompt, which aims to identify meaningful sentences to serve as prompts. Experiments show that ReadPrompt achieves state-of-the-art performance on the current knowledge probing benchmark. Moreover, since the prompt is readable, we discovered a misalignment between constructed prompts and knowledge, which is also present in current prompting methods verified by an attack experiment. We claim that the probing outcomes of the current prompting methods are unreliable that overestimate the knowledge contained within PLMs.

Open-domain dialogue system usually requires different sources of knowledge to generate more informative and evidential responses. However, existing knowledge-grounded dialogue systems either focus on a single knowledge source or overlook the dependency between multiple sources of knowledge, which may result in generating inconsistent or even paradoxical responses. To incorporate multiple knowledge sources and dependencies between them, we propose SAFARI, a novel framework that leverages the exceptional capabilities of large language models (LLMs) in planning, understanding, and incorporating under both supervised and unsupervised settings. Specifically, SAFARI decouples the knowledge grounding into multiple sources and response generation, which allows easy extension to various knowledge sources including the possibility of not using any sources. To study the problem, we construct a personalized knowledge-grounded dialogue dataset Knowledge Behind Persona (KBP), which is the first to consider the dependency between persona and implicit knowledge. Experimental results on the KBP dataset demonstrate that the SAFARI framework can effectively produce persona-consistent and knowledge-enhanced responses.

Recent works usually address Dialog policy learning DPL by training a reinforcement learning (RL) agent to determine the best dialog action. However, existing works on deep RL require a large volume of agent-user interactions to achieve acceptable performance. In this paper, we propose to make full use of the plain text knowledge from the pre-trained language model to accelerate the RL agent’s learning speed. Specifically, we design a dialog action-aware transformer encoder (DaTrans), which integrates a new fine-tuning procedure named masked last action task to encourage DaTrans to be dialog-aware and distill action-specific features. Then, DaTrans is further optimized in an RL setting with ongoing interactions and evolves through exploration in the dialog action space toward maximizing long-term accumulated rewards. The effectiveness and efficiency of the proposed model are demonstrated with both simulator evaluation and human evaluation.