"Kelly is a warm person" paper data

Shorter URL available: kelly.datasette.teddysc.me


The original CSV files are too large, I selected only 100 rows of each file

Full Version

See also the notes on GitHub, including link to paper.


This site is published by Teddy

agency_classifier-agency_bios

202 rows in 3 tables, 505 rows in 15 hidden tables

BIOS_sampled_preprocessed, BIOS_sampled, _source_info_

agency_classifier-agency_dataset

211 rows in 4 tables, 482 rows in 20 hidden tables

train, test, val, _source_info_

biography_dataset-preprocessed_bios

21 tables, 105 hidden tables

df_m_writers_2_para, df_m_sports_2_para, df_m_podcasters_2_para, df_m_musicians_2_para, df_m_models_2_para, ...

biography_dataset-sampled_bios

3 tables, 15 hidden tables

processed_career_life_2_para_df_m, processed_career_life_2_para_df_f, _source_info_

evaluated_letters-chatgpt-cbg

4 tables, 20 hidden tables

all_2_para_w_chatgpt_eval_hallucination_eval, all_2_para_w_chatgpt_eval_hallucination, all_2_para_w_chatgpt_eval, _source_info_

evaluated_letters-chatgpt-clg

101 rows in 2 tables, 282 rows in 10 hidden tables

clg_letters_eval, _source_info_

generated_letters-chatgpt-cbg

23 tables, 115 hidden tables

df_m_writers_2_para_w_chatgpt, df_m_sports_2_para_w_chatgpt, df_m_podcasters_2_para_w_chatgpt, df_m_musicians_2_para_w_chatgpt, df_m_models_2_para_w_chatgpt, ...