Published August 4, 2026 | Version 1.0.0
Dataset Open

AntEngage Empathy Conversation Dataset

Authors/Creators

  • 1. AntEngage Technology Private Limited

Description

A high-quality collection of 4,008 empathy-focused multi-turn conversations generated by AntEngage's AI pipeline and quality-verified through an automated cross-checker. Each conversation simulates realistic emotional support dialogues covering grief, caregiver burnout, loneliness, guilt, workplace exhaustion, and other emotionally complex human experiences.

Created by AntEngage Technology Private Limited as training data for AELM (AntEngage Language Model) — a conversational AI built for India and global markets. All conversations are AI-synthesized with no real user data or PII.

Format: JSONL (primary) and CSV. Each JSONL record contains a conversation_id, language, domain, num_turns, and a list of user/assistant turn pairs.

Files

empathy_conversations.csv

Files (20.0 MB)

Name Size Download all
md5:c82d22230543841e06ab046f9a2ce509
9.6 MB Preview Download
md5:42dbd4648f07bf9c7f69a3386d292e8a
10.5 MB Download

Additional details

Related works

Is documented by
Dataset: https://antengage.com (URL)
Is identical to
Dataset: https://huggingface.co/datasets/antengage/empathy-conversations (URL)
Is supplemented by
Dataset: https://github.com/AntEngage/datasets (URL)

Software

Repository URL
https://github.com/AntEngage/datasets
Development Status
Active