Goal#
target:
1000 unique workplace small talk utterances
scope:
before meeting
after meeting
project-related small talk
week check-in
remote work
team rapport
neutral short answers
Source Plan#
Collection Rules#
do:
collect short neutral workplace utterances
tag opener / answer / follow-up / close
keep source_url and license
filter out travel and consumer-service scenes
deduplicate common greetings
do not:
include private-life-heavy questions
include restaurant / travel / hotel style language
mix generated sentences into the corpus
Corpus Schema#
| id |
utterance |
tag |
source_url |
source_license |
note |