Elevator Small Talk Corpus

Tiny synthetic text snippets labeled by how quickly everyone starts staring at the floor numbers.

⇩ Download

Copy API command

kaggle datasets download -d awkward-nlp/elevator-small-talk-corpus
python - <<'PY'
import pandas as pd
rows = pd.read_csv("corpus/train.jsonl")
print(rows.describe(include="all"))
PY

Elevator Small Talk Corpus3 rows

About Dataset

A harmless fake NLP release for testing text previews, label descriptions, and download affordances in pre-release mocks.

Elevator Small Talk Corpus

train.jsonl (14 KB)

⇩

About this file

Preview file for train.jsonl Preview-sample stats (3 rows).

Autterance_id Link to Organization	123text Text	123label Label	Asplit Split
33 unique · 0% null talk-001 33.3% talk-002 33.3% talk-003 33.3%	33 unique · 0% null Mondays arrive with confidence. 33.3% This elevator has excellent vertical commitment. 33.3% Weather is certainly happening today. 33.3%	33 unique · 0% null calendar 33.3% situational 33.3% weather-neutral 33.3%	22 unique · 0% null train 66.7% validation 33.3%
talk-001	Weather is certainly happening today.	weather-neutral	train
talk-002	This elevator has excellent vertical commitment.	situational	train
talk-003	Mondays arrive with confidence.	calendar	validation

▤

Metadata

☺ShmuggingFace downstream demo (Owner)

Activity Overview

Views34.7K880 in the last 30 days

Downloads404213 in the last 30 days

Comments3posted

Elevator Small Talk Corpus

About Dataset

About this file

See what others are saying about this dataset

What have you used this dataset for?

How would you describe this dataset?

Metadata

Activity Overview