Skip to content

Dataset · Estwld

Empathetic dialogues adapted for LLM training

Estwld/empathetic_dialogues_llm

A cleaned and reformatted version of the Empathetic Dialogues dataset for convenient LLM training.

What's inside

The repository contains between 10,000 and 100,000 English texts from the well-known Empathetic Dialogues dataset. The authors cleaned the data and changed its structure (specifically, combining utterances) so that the dataset can be used directly for training and inference of language models without complex preprocessing.

What you can do with it

  • SFT fine-tuning of chatbots to improve empathy and emotional intelligence
  • Evaluating the ability of language models to maintain a dialogue taking into account the interlocutor's emotions
Tasks
Text generation
Languages
English
Size
10K–100K rows
Downloads this month
1.7K
Likes
33

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("Estwld/empathetic_dialogues_llm")