A subset of the prompts in P3 are still Work in Progress. For information, we provide the lists of the datasets for which prompts have been finalized and datasets for which prompts are suceptible to be modified in the near future. Modifications will in majority consist of metadata collection, but in some cases, will impact the templates themselves.
To facilitate traceability, Promptsource is currently pinned at version 0.1.0
.
Dataset | Subset (optional) |
---|---|
adversarial_qa | dbert |
adversarial_qa | dbidaf |
adversarial_qa | droberta |
adversarial_qa | adversarialQA |
ag_news | |
ai2_arc | ARC-Challenge |
ai2_arc | ARC-Easy |
amazon_polarity | |
anli | |
app_reviews | |
circa | |
cnn_dailymail | 3.0.0 |
common_gen | |
coqa | |
cos_e | v1.11 |
cos_e | v1.0 |
cosmos_qa | |
crows_pairs | |
craffel/openai_lambada | |
dbpedia_14 | |
dream | |
drop | |
duorc | ParaphraseRC |
duorc | SelfRC |
emo | |
gigaword | |
glue | cola |
glue | mrpc |
glue | qqp |
glue | sst2 |
glue | stsb |
hans | |
hellaswag | |
imdb | |
jeopardy | |
jigsaw_toxicity_pred | |
kilt_tasks | nq |
lambada | |
mc_taco | |
multi_news | |
nq_open | |
openbookqa | main |
openbookqa | additional |
paws | labeled_final |
paws | labeled_swap |
paws | unlabeled_final |
paws-x | en |
piqa | |
qa_srl | |
qasc | |
quac | |
quail | |
quarel | |
quartz | |
quoref | |
race | high |
race | middle |
race | all |
ropes | |
rotten_tomatoes | |
samsum | |
sciq | |
scitail | snli_format |
scitail | tsv_format |
social_i_qa | |
squad_v2 | |
super_glue | wsc.fixed |
super_glue | boolq |
super_glue | cb |
super_glue | copa |
super_glue | multirc |
super_glue | record |
super_glue | rte |
super_glue | wic |
swag | regular |
trec | |
trivia_qa | rc |
tydiqa | |
web_questions | |
wiki_bio | |
wiki_hop | original |
wiki_qa | |
winobias | * |
winogender | |
winogrande | winogrande_debiased |
winogrande | winogrande_l |
winogrande | winogrande_m |
winogrande | winogrande_s |
winogrande | winogrande_xl |
winogrande | winogrande_xs |
wiqa | |
xsum | |
yelp_review_full | |
Zaid/coqa_expanded | |
Zaid/quac_expanded |
Dataset | Subset (optional) |
---|---|
acronym_identification | |
ade_corpus_v2 | Ade_corpus_v2_classification |
ade_corpus_v2 | Ade_corpus_v2_drug_ade_relation |
ade_corpus_v2 | Ade_corpus_v2_drug_dosage_relation |
aeslc | |
amazon_reviews_multi | en |
amazon_us_reviews | Wireless_v1_00 |
ambig_qa | light |
aqua_rat | raw |
art | |
asnq | |
asset | ratings |
asset | simplification |
banking77 | |
billsum | |
bing_coronavirus_query_set | |
blended_skill_talk | |
boolq | |
cbt | CN |
cbt | NE |
cbt | P |
cbt | raw |
cbt | V |
cc_news | |
climate_fever | |
codah | codah |
codah | fold_0 |
codah | fold_1 |
codah | fold_2 |
codah | fold_3 |
codah | fold_4 |
commonsense_qa | |
conv_ai | |
conv_ai_2 | |
conv_ai_3 | |
coqa | |
cord19 | metadata |
covid_qa_castorini | |
craigslist_bargains | |
discofuse | discofuse-sport |
discofuse | discofuse-wikipedia |
discovery | discovery |
docred | |
e2e_nlg_cleaned | |
ecthr_cases | alleged-violation-prediction |
emotion | |
esnli | |
evidence_infer_treatment | 1.1 |
evidence_infer_treatment | 2 |
fever | v1.0 |
fever | v2.0 |
financial_phrasebank | sentences_allagree |
freebase_qa | |
generated_reviews_enth | |
glue | ax |
glue | mnli |
glue | mnli_matched |
glue | mnli_mismatched |
glue | qnli |
glue | rte |
glue | wnli |
google_wellformed_query | |
guardian_authorship | cross_genre_1 |
guardian_authorship | cross_topic_1 |
guardian_authorship | cross_topic_4 |
guardian_authorship | cross_topic_7 |
gutenberg_time | |
head_qa | en |
health_fact | |
hlgd | |
hotpot_qa | distractor |
hotpot_qa | fullwiki |
humicroedit | subtask-1 |
humicroedit | subtask-2 |
hyperpartisan_news_detection | byarticle |
hyperpartisan_news_detection | bypublisher |
jfleg | |
kelm | |
liar | |
limit | |
math_dataset | algebra__linear_1d |
math_dataset | algebra__linear_1d_composed |
math_dataset | algebra__linear_2d |
math_dataset | algebra__linear_2d_composed |
math_qa | |
mdd | task1_qa |
mdd | task2_recs |
mdd | task3_qarecs |
medical_questions_pairs | |
meta_woz | dialogues |
mocha | |
movie_rationales | |
multi_nli | |
multi_nli_mismatch | |
multi_x_science_sum | |
mwsc | |
narrativeqa | |
ncbi_disease | |
neural_code_search | evaluation_dataset |
newspop | |
nlu_evaluation_data | |
numer_sense | |
onestop_english | |
poem_sentiment | |
pubmed_qa | pqa_labeled |
qa_zre | |
qed | |
quora | |
samsum | |
scan | addprim_jump |
scan | addprim_turn_left |
scan | filler_num0 |
scan | filler_num1 |
scan | filler_num2 |
scan | filler_num3 |
scan | length |
scan | simple |
scan | template_around_right |
scan | template_jump_around_right |
scan | template_opposite_right |
scan | template_right |
scicite | |
scientific_papers | arxiv |
scientific_papers | pubmed |
scitldr | Abstract |
selqa | answer_selection_analysis |
sem_eval_2010_task_8 | |
sem_eval_2014_task_1 | |
sent_comp | |
sick | |
sms_spam | |
snips_built_in_intents | |
snli | |
species_800 | |
spider | |
squad | |
squad_adversarial | AddSent |
squadshifts | amazon |
squadshifts | new_wiki |
squadshifts | nyt |
sst | default |
stsb_multi_mt | en |
subjqa | books |
subjqa | electronics |
subjqa | grocery |
subjqa | movies |
subjqa | restaurants |
subjqa | tripadvisor |
tab_fact | tab_fact |
tmu_gfm_dataset | |
turk | |
tweet_eval | emoji |
tweet_eval | emotion |
tweet_eval | hate |
tweet_eval | irony |
tweet_eval | offensive |
tweet_eval | sentiment |
tweet_eval | stance_abortion |
tweet_eval | stance_atheism |
tweet_eval | stance_climate |
tweet_eval | stance_feminist |
tweet_eval | stance_hillary |
tydiqa | primary_task |
tydiqa | secondary_task |
wiki_hop | masked |
wiki_qa | |
wiki_split | |
winograd_wsc | wsc273 |
winograd_wsc | wsc285 |
xnli | en |
xquad | xquad.en |
xquad_r | en |
yahoo_answers_qa | |
yahoo_answers_topics | |
yelp_polarity | |
zest |