Dataset Description first test dataset Repository: https://github.com/AsocForDipStudiesAndTraining/ds1/ Point of Contact: info@adst.org or fowlerj@adst.org
Dataset Summary This dataset is wholy comprised of the text transcripts from the Association for Diplomatic Studies and Training's (ADST) Foreign Affairs Oral History Archive.
Languages: The vast majority of the corpus is in the English language (American English). However, due to the nature of the content there are many languages included as words, phrases, or event or place names.
Data Structure
Data Fields author: this is the name of the author interviewer: this is the name of the interviewer transcript: this is the text transcript of the oral history interview
Dataset Creation Curation Rationale: Querying existing LLM models, but limiting their responses to material exsiting in the Foreign Affairs Oral History Archive (FAOHA) was not possible. This dataset was created to fine tune existing models to provide the ability to summerize and query data in the FAOHA.
The source transcripts (all human produced) can be found here: https://adst.org/oral-history/oral-history-interviews
Discussion of Biases These transcripts represent the opinions of individuals. As such they contain many different biases, including historical biases.
Citation Information (to be completed, here for formatting information) @article{article_id, author = {Author Lisa}. title = [Dataset Paper Title], journal = {Publication Venue}, year = {2525} }