The EmpatheticDialogues dataset is a large-scale multi-turn empathetic dialogue dataset collected on the Amazon Mechanical Turk, containing 24,850 one-to-one open-domain conversations. Each conversation was obtained by pairing two crowd-workers: a speaker and a listener. The speaker is asked to talk about the personal emotional feelings. The listener infers the underlying emotion through what the speaker says and responds empathetically. The dataset provides 32 evenly distributed emotion labels.

Empathetic Dialogues for LLM

 This repository contains a reformatted version of the Empathetic Dialogues dataset, tailored for seamless integration with Language Model (LLM) training and inference. The original dataset's format posed challenges for direct application in LLM tasks, prompting us to restructure and clean the data. 

  Data Restructuring

 We have implemented the following changes to enhance the dataset's usability: 

Merged dialogues with the same conv_id… See the full description on the dataset page: https://huggingface.co/datasets/Estwld/empathetic_dialogues_llm.

This dataset is about books. It has 1 row and is filtered where the book is Beyond the walls : Abraham Joshua Heschel and Edith Stein on the significance of empathy for Jewish-Christian dialogue. It features 7 columns including author, publication date, language, and book publisher.

Social interaction has been recognized as positively affecting learning, with dialogue–as a common form of social interaction–comprising an integral part of collaborative learning. Interactive storytelling is defined as a branching narrative in which users can experience different story lines with alternative endings, depending on the choices they make at various decision points of the story plot. In this research, we aim to harness the power of dialogic practices by incorporating dialogic activities in the decision points of interactive digital storytelling experiences set in a history education context. Our objective is to explore interactive storytelling as a collaborative learning experience for remote learners, as well as its effect on promoting historical empathy. As a preliminary validation of this concept, we recorded the perspective of 14 educators, who supported the value of the specific conceptual design. Then, we recruited 15 adolescents who participated in our main study in 6 groups. They were called to experience collaboratively an interactive storytelling experience set in the Athens Ancient Agora (Market) wherein we used the story decision/branching points as incentives for dialogue. Our results suggest that this experience design can indeed support small groups of remote users, in-line with special circumstances like those of the COVID-19 pandemic, and confirm the efficacy of the approach to establish engagement and promote affect and reflection on historical content. Our contribution thus lies in proposing and validating the application of interactive digital storytelling as a dialogue-based collaborative learning experience for the education of history.

This research aims to develop an in-depth understanding of empathic perspective-taking in design teaching. A pedagogical probe study was utilised to obtain design teachers’ insights into adopting allocentric views by taking the perspectives of design students, users, and materials during design conversations. This participatory research study enables design teachers and supporting teaching staff to contribute to the research process by self-documenting actively and reflecting on their teaching approaches. The findings show that the perspective-taking of design teachers relies on distinctive pedagogical intentions, cultivated from intrinsic concerns and personal motivations, that can spark motivation and inspiration in the design learning processes of students.

This dataset is the Multilingual Corpus of the DIALLS (DIalogue and Argumentation for Literacy Learning in Schools) project (dialls2020.eu/) and consists of a set of transcripts of classroom interactions of students from ages 5 to 15 years old. These classroom interactions took place in seven DIALLS participant countries (UK, Portugal, Germany, Lithuania, Spain, Cyprus, and Israel).

The corpus is a set of 202 transcripts in the participant countries’ native language (English, Portuguese, German, Lithuanian, Catalan, Cypriot Greek, and Hebrew). The transcripts in each native language range from a maximum of 35 for Hebrew (more than 17% of the overall corpus) to a minimum of 19 transcripts for Cypriot Greek (10% of the corpus). More than 50% of the transcripts in a language different from English (90 transcripts) have associated their English translation. The topic of the project is cultural literacy through dialogue and argumentation in school children.

The Multilingual Corpus is relevant to the following areas of research: Educational dialogue, Citizenship education, Argumentation and learning, Multimodal literacy, Dialogic teaching, Dialogue/discourse analysis, Arts-based education, Cultural studies, Teacher professional development and communities.

The dataset is organised in two main sections: an Excel file, and a zip folder with .csv files matching the excel file. A description of the corpus and further information on the dataset can be accessed in a .pdf file.

This study describes a preliminary test of the hypothesis that, when people engage in musical and linguistic activities designed to enhance the interactive, turn-taking properties of typical conversation, they benefit in ways that enhance empathy and self-esteem, relative to people who experience activities that are similar except that synchronous action is emphasized, with no interactional turn-taking. Twenty-two 12–14 year olds identified as socially vulnerable (e.g., for anxiety) received six enjoyable 1-h sessions of musical improvisation, language games that developed sensitivity to linguistic rhythm and melody, and cross-over activities like rap. The Turn-taking group (n = 11), practiced characteristics of conversation in language games, and these were also introduced into musical activities. This involved much turn-taking and predicting what others would do. A matched control group, the Synchrony group, did similar activities but in synchrony, with less prediction and no turn-taking. Task complexity increased over the six sessions. Psychometric testing before and after the series showed that the Turn-taking group increased in empathy on self-report (Toronto Empathy Questionnaire) and behavioral (‘Reading the Mind in the Eyes’) measures, and in the General subtest of the Culture-Free Self-Esteem Inventory. While more work is needed to confirm the conclusions for relevant demographic groups, the current results point to the social value of musical and linguistic activities that mimic entrained, tightly coordinated parameters of everyday conversational interaction, in which, at any one time, individuals act as equal participants who have different roles.

CCI Dataset V2

  Dataset Summary

The CCI (Customer Conversation Intelligence) Dataset V2 is a comprehensive collection of airline customer service interactions from Intelligent Virtual Assistant (IVA) conversations. The dataset contains 15,598 entries with rich annotations including intent prediction, sentiment analysis, empathy scoring, and conversation topic classification. Originally sourced from airline customer service interactions, this dataset provides valuable… See the full description on the dataset page: https://huggingface.co/datasets/raghavdw/cci-dataset-v2.

Classroom discussions of current events and controversial topics can devolve into unproductive and highly charged debates. This article describes an in-class exercise used to foster respect during difficult conversations, by encouraging students to design rules for discussions and guidelines for creating a safe space for dialogue. This activity relies on three underlying principles: trust, empowerment and empathy. These principles can be integrated into a broader pedagogical approach, which emphasizes a democratic classroom and active learning. Student feedback shows that the intervention can be useful for promoting respectful and engaging discussions during moments of tension and polarization. But an emphasis on civility may also undermine the diversity of opinions, and require respecting student silences.

Utilized communications skills.

Participant workshop evaluations.

Large language models (LLMs) have shown promise for task-oriented dialogue across a range of domains. The use of LLMs in health and fitness coaching is under-explored. Behavior science frameworks such as COM-B, which conceptualizes behavior change in terms of capability (C), Opportunity (O) and Motivation (M), can be used to architect coaching interventions in a way that promotes sustained change. Here we aim to incorporate behavior science principles into an LLM using two knowledge infusion techniques: coach message priming (where exemplar coach responses are provided as context to the LLM), and dialogue re-ranking (where the COM-B category of the LLM output is matched to the inferred user need). Simulated conversations were conducted between the primed or unprimed LLM and a member of the research team, and then evaluated by 8 human raters. Ratings for the primed conversations were significantly higher in terms of empathy and actionability. The same raters also compared a single response generated by the unprimed, primed and re-ranked models, finding a significant uplift in actionability and empathy from the re-ranking technique. This is a proof of concept of how behavior science frameworks can be infused into automated conversational agents for a more principled coaching experience.