Datasets:
metadata
pretty_name: MathMentorDB sample (50 conversations, ungated)
license: cc-by-nc-sa-4.0
language:
- en
task_categories:
- text-classification
tags:
- education
- tutoring
- mathematics
- dialogue
size_categories:
- n<1K
MathMentorDB sample (ungated)
An ungated 50-conversation sample of the MathMentorDB labeled subset, provided
so the data can be inspected without an access request. The full labeled
subset (7,000 conversations, 165,275 messages, 24-move discourse labels and
conversation resolution labels) is available gated under a Data Use Agreement
at mathmentordb/MathMentorDB,
whose dataset card documents the schema, taxonomy, provenance, and ethics
considerations. This sample uses the identical schema
(data/sample.jsonl, one conversation per line).