Replication Data for: Synthetically generated text for supervised text analysis (doi:10.7910/DVN/JJ5BBX)

View:

Part 1: Document Description
Part 2: Study Description
Part 3: Data Files Description
Part 4: Variable Description
Part 5: Other Study-Related Materials
Entire Codebook

Replication Data for: Synthetically generated text for supervised text analysis

doi:10.7910/DVN/JJ5BBX

Harvard Dataverse

2024-11-15

1

Halterman, Andrew, 2024, "Replication Data for: Synthetically generated text for supervised text analysis", https://doi.org/10.7910/DVN/JJ5BBX, Harvard Dataverse, V1, UNF:6:JJUrUpeMWFKHndQZmjKvEw== [fileUNF]

Replication Data for: Synthetically generated text for supervised text analysis

doi:10.7910/DVN/JJ5BBX

Halterman, Andrew (Michigan State University)

<i>Political Analysis</i>

Harvard Dataverse

Halterman, Andrew

Halterman, Andrew

2024-09-25

https://doi.org/10.7910/DVN/JJ5BBX

Social Sciences

Large language models are a powerful tool for conducting text analysis in political science, but using them to annotate text has several drawbacks, including high cost, limited reproducibility, and poor explainability. Traditional supervised text classifiers are fast and reproducible, but require expensive hand annotation, which is especially difficult for rare classes. This article proposes using LLMs to generate synthetic training data for training smaller, traditional supervised text models. Synthetic data can augment limited hand annotated data or be used on its own to train a classifier with good performance and greatly reduced cost. I provide a conceptual overview of text generation, guidance on when researchers should prefer different techniques for generating synthetic text, a discussion of ethics, a simple technique for improving the quality of synthetic text, and an illustration of its limitations. I demonstrate the usefulness of synthetic training through three validations: synthetic news articles describing police responses to communal violence in India for training an event detection system, a multilingual corpus of synthetic populist manifesto statements for training a sentence-level populism classifier, and generating synthetic tweets describing the fighting in Ukraine to improve a named entity system.

Forthcoming, Political Analysis

Forthcoming, Political Analysis

• Number of cases: 900

• No. of variables per record: 3

• Type of File: text/tab-separated-values

UNF:6:o1yVfXB2c0gFEn41HHqH+g==

• Number of cases: 780

• No. of variables per record: 3

• Type of File: text/tab-separated-values

UNF:6:aqyDEjVF3ucBtr+Ulx4fqw==

• Number of cases: 15000

• No. of variables per record: 5

• Type of File: text/tab-separated-values

UNF:6:lK9pYarmAGs5dUmaGSTcsQ==

• Number of cases: 15000

• No. of variables per record: 11

• Type of File: text/tab-separated-values

UNF:6:T3PDuksi+oYhlHqF7l4WTw==

• Number of cases: 20400

• No. of variables per record: 11

• Type of File: text/tab-separated-values

UNF:6:mOEWqLvcz1OO4RTHNymJvw==

• Number of cases: 252

• No. of variables per record: 4

• Type of File: text/tab-separated-values

UNF:6:FOpz1zA+ucO3K+sfppK6fA==

• text - text
• title - title
• label - label
• text - text
• title - title
• label - label
• num_annot - num_annot
• f1 - f1
• event_type - event_type
• seed - seed
• method - method
• num_annot - num_annot
• f1 - f1
• precision - precision
• recall - recall
• accuracy - accuracy
• remaining_pos - remaining_pos
• perc_positives - perc_positives
• event_type - event_type
• class_weight - class_weight
• seed - seed
• method - method
• n_human - n_human
• n_synth_arg - n_synth_arg
• seed - seed
• synth_type - synth_type
• f1 - f1
• perc_pos - perc_pos
• synth_n - synth_n
• event_type - event_type
• traditional_augmentation - traditional_augmentation
• balanced_class_weight - balanced_class_weight
• model - model
• token_f1 - token_f1
• n - n
• train_file - train_file
• seed - seed

f10551855 Location:

Variable Format: character

Notes: UNF:6:78CTScnIupbV/iwRTN3F0g==

f10551855 Location:

Variable Format: character

Notes: UNF:6:gctIOeV7pWlbwIQjm88CiQ==

f10551855 Location:

Variable Format: character

Notes: UNF:6:1R3P7ynfEnxye2gFSONclg==

f10551861 Location:

Variable Format: character

Notes: UNF:6:RDYIwuD2LF0i1ZR8zgUfOA==

f10551861 Location:

Variable Format: character

Notes: UNF:6:OjPDJv6HEWDRTPvaVC4fPA==

f10551861 Location:

Variable Format: character

Notes: UNF:6:m4NJ/YD4elCDVlcPgdN+QA==

f10551866 Location:

Summary Statistics: Valid 15000.0; Max. 980.0; Mean 490.0; StDev 288.6270148540851; Min. 0.0

Variable Format: numeric

Notes: UNF:6:kqoT4F+MeRjkMynONRezJQ==

f10551866 Location:

Summary Statistics: Max. 0.7710843373493976; StDev 0.23284022831277068; Mean 0.4988129145868635; Valid 15000.0; Min. 0.0

Variable Format: numeric

Notes: UNF:6:X5X9be5xn0LUcArY6tFj1Q==

f10551866 Location:

Variable Format: character

Notes: UNF:6:OsTFQCjxH53pFaeVQW/Itg==

f10551866 Location:

Summary Statistics: Min. 0.0; Max. 49.0; Valid 15000.0; Mean 24.5; StDev 14.431350742704252

Variable Format: numeric

Notes: UNF:6:gSjV0bXJpCYbHHTbUBYEXw==

f10551866 Location:

Variable Format: character

Notes: UNF:6:0dU3gT/6AOWw+austuV1/A==

f10551856 Location:

Summary Statistics: Valid 15000.0; Max. 980.0; Mean 490.0; StDev 288.6270148540851; Min. 0.0

Variable Format: numeric

Notes: UNF:6:kqoT4F+MeRjkMynONRezJQ==

f10551856 Location:

Summary Statistics: Valid 15000.0; Max. 0.7631578947368421; StDev 0.22581957122092397; Min. 0.0; Mean 0.5056345091694852;

Variable Format: numeric

Notes: UNF:6:FSbCFxrP0zvFu350oOcwCQ==

f10551856 Location:

Summary Statistics: Mean 0.7305913603284444; Valid 15000.0; Max. 1.0; StDev 0.17460428030540925; Min. 0.0

Variable Format: numeric

Notes: UNF:6:QiLZIlN6PC5zFSOihgoYIg==

f10551856 Location:

Summary Statistics: Valid 15000.0; StDev 0.21691624152795394; Mean 0.4188488816738817; Min. 0.0; Max. 0.79375

Variable Format: numeric

Notes: UNF:6:n88JmV8S76+/s45UnEsa7g==

f10551856 Location:

Summary Statistics: Min. 0.931469708302169; Max. 0.9983171278982798; StDev 0.0037931706497809753; Mean 0.9931895350286711; Valid 15000.0

Variable Format: numeric

Notes: UNF:6:wnLIJhQfN6HL/PfmIKEgZA==

f10551856 Location:

Summary Statistics: Mean 34.02846666666669; Max. 139.0; Valid 15000.0; StDev 30.183045860207248; Min. 0.0;

Variable Format: numeric

Notes: UNF:6:tpPGcupHRTcZHAfz64mwkg==

f10551856 Location:

Summary Statistics: StDev 0.08109927876184747; Max. 0.6666666666666666; Mean 0.15640469913651195; Min. 0.005263157894736842; Valid 15000.0

Variable Format: numeric

Notes: UNF:6:Gmgy+P8bC4LX9RVmYrRF/Q==

f10551856 Location:

Variable Format: character

Notes: UNF:6:OsTFQCjxH53pFaeVQW/Itg==

f10551856 Location:

Summary Statistics: Valid 0.0; Mean NaN; Min. NaN; StDev NaN; Max. NaN

Variable Format: numeric

Notes: UNF:6:GjAafq4oaAd+hOiWnQmCvQ==

f10551856 Location:

Summary Statistics: Valid 15000.0; Max. 49.0; StDev 14.431350742704252; Mean 24.5; Min. 0.0;

Variable Format: numeric

Notes: UNF:6:gSjV0bXJpCYbHHTbUBYEXw==

f10551856 Location:

Variable Format: character

Notes: UNF:6:0dU3gT/6AOWw+austuV1/A==

f10551864 Location:

Summary Statistics: Mean 470.5882352941176; Max. 1000.0; Min. 0.0; StDev 386.1894081819535; Valid 20400.0

Variable Format: numeric

Notes: UNF:6:Ms8vXs/wAVL8xX4RAxRYEQ==

f10551864 Location:

Summary Statistics: Max. 500.0; Valid 20400.0; StDev 186.29643865896256; Min. -1.0; Mean 164.5

Variable Format: numeric

Notes: UNF:6:4GMWskCDWDoK4G/Gg9s/xg==

f10551864 Location:

Summary Statistics: Min. 0.0; Max. 49.0; Valid 20400.0; StDev 14.43122340045245; Mean 24.5

Variable Format: numeric

Notes: UNF:6:cdDmDydaQplZrwSW0Lp5Og==

f10551864 Location:

Variable Format: character

Notes: UNF:6:kMRvXsTRDJELlIL2So0ghw==

f10551864 Location:

Summary Statistics: Max. 0.7220216606498194; Valid 20400.0; StDev 0.2096684957771509; Min. 0.0; Mean 0.37664428997884664

Variable Format: numeric

Notes: UNF:6:0WgODARLHZLb+JfrrMbkMw==

f10551864 Location:

Summary Statistics: Min. 0.0; Max. 0.31; StDev 0.06232873741633384; Valid 20400.0; Mean 0.06542067316681392

Variable Format: numeric

Notes: UNF:6:gEJTll7xbSZOAJOl//cFsw==

f10551864 Location:

Summary Statistics: Min. 0.0; Valid 20400.0; Mean 430.29411764705884; Max. 1680.0; StDev 500.5471899352906;

Variable Format: numeric

Notes: UNF:6:eJe+l0qZWSNGI5jkEW44yQ==

f10551864 Location:

Variable Format: character

Notes: UNF:6:q32hJL6/r5/fUQK72SvuIA==

f10551864 Location:

Variable Format: character

Notes: UNF:6:KAv9S2dcKNPGVwdqSkcamA==

f10551864 Location:

Variable Format: character

Notes: UNF:6:EEVgM2SxZLA1tT3mukKJww==

f10551864 Location:

Variable Format: character

Notes: UNF:6:qS8ULz43CDIXOygTm9JbkA==

f10552042 Location:

Summary Statistics: Mean 0.5133968216481106; Max. 0.7792553191489362; Min. 0.0; StDev 0.20413850170125777; Valid 252.0;

Variable Format: numeric

Notes: UNF:6:jZCU8FXT1c+f6WRwmCv95A==

f10552042 Location:

Summary Statistics: Valid 252.0; Max. 300.0; StDev 68.7989030868835; Min. 50.0; Mean 127.77777777777773

Variable Format: numeric

Notes: UNF:6:aCyG/DAq6fjqOWYT2Z62nQ==

f10552042 Location:

Variable Format: character

Notes: UNF:6:XsUQFWhx5J8/Smk8kyIRgg==

f10552042 Location:

Summary Statistics: Min. 1.0; Max. 14.0; Mean 7.5; StDev 4.039151029097151; Valid 252.0;

Variable Format: numeric

Notes: UNF:6:yWxTbHu/0pCTK1/Gleu8bw==

Label:

README.txt

text/plain

Label:

requirements.txt

text/plain

Label:

README.md

Text:

text/markdown

Label:

raw_annotations--raw.jsonl

Text:

application/octet-stream

Label:

sents.csv

Text:

text/comma-separated-values

Label:

.DS_Store

Text:

application/octet-stream

Label:

.DS_Store

Text:

application/octet-stream

Label:

embedding_vis.py

Text:

text/x-python-script

Label:

generate_figures.pdf

Text:

application/pdf

Label:

generate_figures.Rmd

Text:

text/x-r-notebook

Label:

gen_synth_india.py

Text:

text/x-python-script

Label:

india_police_events_active_learning.py

Text:

text/x-python-script

Label:

india_police_events_experiment.py

Text:

text/x-python-script

Label:

theme_pub.R

Text:

type/x-r-syntax

Label:

IPE_events_active_error.png

Text:

image/png

Label:

IPE_synth_diff.png

Text:

image/png

Label:

new_ipe_fig_synth_both.png

Text:

image/png

Label:

new_ipe_fig_synth_unbalanced.png

Text:

image/png

Label:

new_ipe_fig_unbalanced.png

Text:

image/png

Label:

.DS_Store

Text:

application/octet-stream

Label:

README.md

Text:

text/markdown

Label:

.DS_Store

Text:

application/octet-stream

Label:

cmp_labeled_sentences_lang.csv

Text:

text/comma-separated-values

Label:

cmp_labeled_synth_statements_2024_02_16.jsonl

Text:

application/octet-stream

Label:

cmp_labeled_synth_statements_2024_02_18.jsonl

Text:

application/octet-stream

Label:

gpt3_synth_all_cmp_2022-12-15.jsonl

Text:

application/octet-stream

Label:

gpt3_synth_populism2.jsonl

Text:

application/octet-stream

Label:

gpt3_synth_populism_neg2.jsonl

Text:

application/octet-stream

Label:

manifesto_sent_level.jsonl

Text:

application/octet-stream

Label:

manifesto_sent_level_for_prodigy.jsonl

Text:

application/octet-stream

Label:

manifesto_sent_level_score_2023-02-05.csv

Text:

text/comma-separated-values

Label:

manifesto_sent_level_score_mlm.csv

Text:

text/comma-separated-values

Label:

populism_hand_validation.csv

Text:

text/comma-separated-values

Label:

populism_validation.jsonl

Text:

application/octet-stream

Label:

fit_cmp_classifier_new.py

Text:

text/x-python-script

Label:

generate_cmp_statements.py

Text:

text/x-python-script

Label:

gen_synth_populist.py

Text:

text/x-python-script

Label:

get_top_ukip.py

Text:

text/x-python-script

Label:

hand_validation.py

Text:

text/x-python-script

Label:

synth_pop_classifier_setfit.py

Text:

text/x-python-script

Label:

theme_pub.R

Text:

type/x-r-syntax

Label:

config.json

Text:

application/json

Label:

config_sentence_transformers.json

Text:

application/json

Label:

config_setfit.json

Text:

application/json

Label:

model.safetensors

Text:

application/octet-stream

Label:

model_head.pkl

Text:

application/octet-stream

Label:

modules.json

Text:

application/json

Label:

pytorch_model.bin

Text:

application/macbinary

Label:

README.md

Text:

text/markdown

Label:

sentencepiece.bpe.model

Text:

application/octet-stream

Label:

sentence_bert_config.json

Text:

application/json

Label:

special_tokens_map.json

Text:

application/json

Label:

tokenizer.json

Text:

application/json

Label:

tokenizer_config.json

Text:

application/json

Label:

config.json

Text:

application/json

Label:

config.json

Text:

application/json

Label:

config_sentence_transformers.json

Text:

application/json

Label:

model_head.pkl

Text:

application/octet-stream

Label:

modules.json

Text:

application/json

Label:

pytorch_model.bin

Text:

application/macbinary

Label:

README.md

Text:

text/markdown

Label:

sentencepiece.bpe.model

Text:

application/octet-stream

Label:

sentence_bert_config.json

Text:

application/json

Label:

special_tokens_map.json

Text:

application/json

Label:

tokenizer.json

Text:

application/json

Label:

tokenizer_config.json

Text:

application/json

Label:

config.json

Text:

application/json

Label:

README.md

Text:

text/markdown

Label:

tweet_text_partial.csv

Text:

text/comma-separated-values

Label:

fake_ukraine_tweets_2022-07-25_0.8_20_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_20_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_20_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_20_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_20_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_20_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_20_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_50_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_50_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_50_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_50_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_50_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_50_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_50_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_80_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_80_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_80_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_80_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_80_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_80_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.8_80_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_20_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_20_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_20_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_20_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_20_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_20_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_20_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_50_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_50_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_50_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_50_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_50_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_50_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_50_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_80_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_80_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_80_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_80_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_80_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_80_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.95_80_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_20_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_20_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_20_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_20_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_20_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_20_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_20_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_50_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_50_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_50_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_50_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_50_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_50_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_50_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_80_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_80_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_80_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_80_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_80_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_80_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.99_80_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_20_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_20_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_20_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_20_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_20_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_20_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_20_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_50_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_50_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_50_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_50_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_50_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_50_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_50_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_80_0.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_80_0.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_80_0.7_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_80_1.3_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_80_1.5_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_80_1.8_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-25_0.9_80_1_3.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.8_50_0.3_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.8_50_0.5_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.8_50_0.7_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.8_50_1.3_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.8_50_1.5_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.8_50_1.8_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.8_50_1_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.95_50_0.3_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.95_50_0.5_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.95_50_0.7_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.95_50_1.3_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.95_50_1.5_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.95_50_1.8_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.95_50_1_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.99_50_0.3_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.99_50_0.5_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.99_50_0.7_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.99_50_1.3_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.99_50_1.5_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.99_50_1.8_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.99_50_1_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.9_50_0.3_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.9_50_0.5_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.9_50_0.7_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.9_50_1.3_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.9_50_1.5_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.9_50_1.8_1.jsonl

Text:

application/octet-stream

Label:

fake_ukraine_tweets_2022-07-26_0.9_50_1_1.jsonl

Text:

application/octet-stream

Label:

synth_detection_results_transformer.csv

Text:

text/comma-separated-values

Label:

ukr_synth_bad_both.jsonl

Text:

application/octet-stream

Label:

ukr_weapon_real.jsonl

Text:

application/octet-stream

Label:

ukr_weapon_synth_best3.jsonl

Text:

application/octet-stream

Label:

1_get_tweets.py

Text:

text/x-python-script

Label:

2_fine_tune_gpt2_twitter.py

Text:

text/x-python-script

Label:

3_generate_synth_tweets_gpt2.py

Text:

text/x-python-script

Label:

4_select_generation_parameters_fancy.py

Text:

text/x-python-script

Label:

5_fit_ner.py

Text:

text/x-python-script

Label:

6_real_vs_synth_ner_plots.Rmd

text/x-r-notebook

Label:

theme_pub.R

Text:

type/x-r-syntax

Label:

real_vs_synth_ner_perf.pdf

application/pdf

Label:

tweet_discrim_parameters_transformer.pdf

Text:

application/pdf

Label:

config.json

Text:

application/json

Label:

generation_config.json

Text:

application/json

Label:

model.safetensors

application/octet-stream

Label:

training_args.bin

Text:

application/macbinary