Interpretability tools that offer explanations in the form of a dialogue have demonstrated their efficacy in enhancing users’ understanding (Slack et al., 2023; Shen et al., 2023), as one-off explanations may fall short in providing sufficient information to the user. Current solutions for dialogue-based explanations, however, often require external tools and modules and are not easily transferable to tasks they were not designed for. With LLMCheckup, we present an easily accessible tool that allows users to chat with any state-of-the-art large language model (LLM) about its behavior. We enable LLMs to generate explanations and perform user intent recognition without fine-tuning, by connecting them with a broad spectrum of Explainable AI (XAI) methods, including white-box explainability tools such as feature attributions, and self-explanations (e.g., for rationale generation). LLM-based (self-)explanations are presented as an interactive dialogue that supports follow-up questions and generates suggestions. LLMCheckup provides tutorials for operations available in the system, catering to individuals with varying levels of expertise in XAI and supporting multiple input modalities. We introduce a new parsing strategy that substantially enhances the user intent recognition accuracy of the LLM. Finally, we showcase LLMCheckup for the tasks of fact checking and commonsense question answering. Our code repository: https://github.com/DFKI-NLP/LLMCheckup

Conversational explainable artificial intelligence (ConvXAI) systems based on large language models (LLMs) have garnered significant interest from the research community in natural language processing (NLP) and human-computer interaction (HCI). Such systems can provide answers to user questions about explanations in dialogues, have the potential to enhance users’ comprehension and offer more information about the decision-making and generation processes of LLMs. Currently available ConvXAI systems are based on intent recognition rather than free chat, as this has been found to be more precise and reliable in identifying users’ intentions. However, the recognition of intents still presents a challenge in the case of ConvXAI, since little training data exist and the domain is highly specific, as there is a broad range of XAI methods to map requests onto. In order to bridge this gap, we present CoXQL, the first dataset in the NLP domain for user intent recognition in ConvXAI, covering 31 intents, seven of which require filling multiple slots. Subsequently, we enhance an existing parsing approach by incorporating template validations, and conduct an evaluation of several LLMs on CoXQL using different parsing strategies. We conclude that the improved parsing approach (MP+) surpasses the performance of previous approaches. We also discover that intents with multiple slots remain highly challenging for LLMs.

Misunderstandings occur all the time in human conversation but deciding on when to ask for clarification is a challenging task for conversational systems that requires a balance between asking too many unnecessary questions and running the risk of providing incorrect information. This work investigates clarification identification based on the task and data from (Xu et al., 2019), reproducing their Transformer baseline and extending it by comparing pre-trained language model fine-tuning, prompt tuning and manual prompt engineering on the task of clarification identification. Our experiments show strong performance with LM and a prompt tuning approach with BERT and RoBERTa, outperforming standard LM fine-tuning, while manual prompt engineering with GPT-3.5 proved to be less effective, although informative prompt instructions have the potential of steering the model towards generating more accurate explanations for why clarification is needed.

This paper presents the dfki-mlst submission for the DialAM shared task (Ruiz-Dolz et al., 2024) on identification of argumentative and illocutionary relations in dialogue. Our model achieves best results in the global setting: 48.25 F1 at the focused level when looking only at the related arguments/locutions and 67.05 F1 at the general level when evaluating the complete argument maps. We describe our implementation of the data pre-processing, relation encoding and classification, evaluating 11 different base models and performing experiments with, e.g., node text combination and data augmentation. Our source code is publicly available.

In this paper we describe the task of adapting NLP models to dialogue processing in the emergency response domain. Our goal is to provide a recipe for building a system that performs dialogue act classification and domain-specific slot tagging while being efficient, flexible and robust. We show that adapter models Pfeiffer et al. (2020) perform well in the emergency response domain and benefit from additional dialogue context and speaker information. Comparing adapters to standard fine-tuned Transformer models we show that they achieve competitive results and can easily accommodate new tasks without significant memory increase since the base model can be shared between the adapters specializing on different tasks. We also address the problem of scarce annotations in the emergency response domain and evaluate different data augmentation techniques in a low-resource setting.

While recently developed NLP explainability methods let us open the black box in various ways (Madsen et al., 2022), a missing ingredient in this endeavor is an interactive tool offering a conversational interface. Such a dialogue system can help users explore datasets and models with explanations in a contextualized manner, e.g. via clarification or follow-up questions, and through a natural language interface. We adapt the conversational explanation framework TalkToModel (Slack et al., 2022) to the NLP domain, add new NLP-specific operations such as free-text rationalization, and illustrate its generalizability on three NLP tasks (dialogue act classification, question answering, hate speech detection). To recognize user queries for explanations, we evaluate fine-tuned and few-shot prompting models and implement a novel adapter-based approach. We then conduct two user studies on (1) the perceived correctness and helpfulness of the dialogues, and (2) the simulatability, i.e. how objectively helpful dialogical explanations are for humans in figuring out the model’s predicted label when it’s not shown. We found rationalization and feature attribution were helpful in explaining the model behavior. Moreover, users could more reliably predict the model outcome based on an explanation dialogue rather than one-off explanations.

The paper presents two multilingual coreference resolution systems submitted for the CRAC Shared Task 2023. The DFKI-Adapt system achieves 61.86 F1 score on the shared task test data, outperforming the official baseline by 4.9 F1 points. This system uses a combination of different features and training settings, including character embeddings, adapter modules, joint pre-training and loss-based re-training. We provide evaluation for each of the settings on 12 different datasets and compare the results. The other submission DFKI-MPrompt uses a novel approach that involves prompting for mention generation. Although the scores achieved by this model are lower compared to the baseline, the method shows a new way of approaching the coreference task and provides good results with just five epochs of training.

We describe three models submitted for the CODI-CRAC 2022 shared task. To perform identity anaphora resolution, we test several combinations of the incremental clustering approach based on the Workspace Coreference System (WCS) with other coreference models. The best result is achieved by adding the “cluster merging” version of the coref-hoi model, which brings up to 10.33% improvement1 over vanilla WCS clustering. Discourse deixis resolution is implemented as multi-task learning: we combine the learning objective of coref-hoi with anaphor type classification. We adapt the higher-order resolution model introduced in Joshi et al. (2019) for bridging resolution given gold mentions and anaphors.

We describe the system developed by the DFKI-TalkingRobots Team for the CODI-CRAC 2021 Shared-Task on anaphora resolution in dialogue. Our system consists of three subsystems: (1) the Workspace Coreference System (WCS) incrementally clusters mentions using semantic similarity based on embeddings combined with lexical feature heuristics; (2) the Mention-to-Mention (M2M) coreference resolution system pairs same entity mentions; (3) the Discourse Deixis Resolution (DDR) system employs a Siamese Network to detect discourse anaphor-antecedent pairs. WCS achieved F1-score of 55.6% averaged across the evaluation test sets, M2M achieved 57.2% and DDR achieved 21.5%.

We compare our team’s systems to others submitted for the CODI-CRAC 2021 Shared-Task on anaphora resolution in dialogue. We analyse the architectures and performance, report some problematic cases in gold annotations, and suggest possible improvements of the systems, their evaluation, data annotation, and the organization of the shared task.

This work addresses coreference resolution in Abstract Meaning Representation (AMR) graphs, a popular formalism for semantic parsing. We evaluate several current coreference resolution techniques on a recently published AMR coreference corpus, establishing baselines for future work. We also demonstrate that coreference resolution can improve the accuracy of a state-of-the-art semantic parser on this corpus.

We present the results we obtained on the classification of dialogue acts in a corpus of human-human team communication in the domain of robot-assisted disaster response. We annotated dialogue acts according to the ISO 24617-2 standard scheme and carried out experiments using the FastText linear classifier as well as several neural architectures, including feed-forward, recurrent and convolutional neural models with different types of embeddings, context and attention mechanism. The best performance was achieved with a ”Divide & Merge” architecture presented in the paper, using trainable GloVe embeddings and a structured dialogue history. This model learns from the current utterance and the preceding context separately and then combines the two generated representations. Average accuracy of 10-fold cross-validation is 79.8%, F-score 71.8%.