Datasets:

Modalities:
Text
Formats:
parquet
Languages:
English
Libraries:
Datasets
pandas
License:
seq
stringlengths
14
66
label
stringclasses
2 values
Adj_Class
stringclasses
3 values
Adj
stringlengths
2
14
Nn
stringlengths
1
17
Hy
stringlengths
2
19
A soviet assertions is a assertions
True
I
soviet
assertions
declaration
A soviet assertions is a declaration
True
I
soviet
assertions
declaration
A soviet assertions is a soviet declaration
True
I
soviet
assertions
declaration
A financial assertions is a assertions
True
N
financial
assertions
declaration
A financial assertions is a declaration
True
N
financial
assertions
declaration
A financial assertions is a financial declaration
False
N
financial
assertions
declaration
A debatable assertions is a assertions
False
O
debatable
assertions
declaration
A debatable assertions is a declaration
False
O
debatable
assertions
declaration
A debatable assertions is a debatable declaration
True
O
debatable
assertions
declaration
A phony assertions is a assertions
False
O
phony
assertions
declaration
A phony assertions is a declaration
False
O
phony
assertions
declaration
A phony assertions is a phony declaration
True
O
phony
assertions
declaration
A hypothetical assertions is a assertions
False
O
hypothetical
assertions
declaration
A hypothetical assertions is a declaration
False
O
hypothetical
assertions
declaration
A hypothetical assertions is a hypothetical declaration
True
O
hypothetical
assertions
declaration
A erroneous assertions is a assertions
False
O
erroneous
assertions
declaration
A erroneous assertions is a declaration
False
O
erroneous
assertions
declaration
A erroneous assertions is a erroneous declaration
True
O
erroneous
assertions
declaration
A portuguese animosity is a animosity
True
I
portuguese
animosity
hostility
A portuguese animosity is a hostility
True
I
portuguese
animosity
hostility
A portuguese animosity is a portuguese hostility
True
I
portuguese
animosity
hostility
A substantial animosity is a animosity
True
N
substantial
animosity
hostility
A substantial animosity is a hostility
True
N
substantial
animosity
hostility
A substantial animosity is a substantial hostility
False
N
substantial
animosity
hostility
A former animosity is a animosity
False
O
former
animosity
hostility
A former animosity is a hostility
False
O
former
animosity
hostility
A former animosity is a former hostility
True
O
former
animosity
hostility
A apparent animosity is a animosity
False
O
apparent
animosity
hostility
A apparent animosity is a hostility
False
O
apparent
animosity
hostility
A apparent animosity is a apparent hostility
True
O
apparent
animosity
hostility
A past animosity is a animosity
False
O
past
animosity
hostility
A past animosity is a hostility
False
O
past
animosity
hostility
A past animosity is a past hostility
True
O
past
animosity
hostility
A alleged animosity is a animosity
False
O
alleged
animosity
hostility
A alleged animosity is a hostility
False
O
alleged
animosity
hostility
A alleged animosity is a alleged hostility
True
O
alleged
animosity
hostility
A western constellations is a constellations
True
I
western
constellations
design
A western constellations is a design
True
I
western
constellations
design
A western constellations is a western design
True
I
western
constellations
design
A notable constellations is a constellations
True
N
notable
constellations
design
A notable constellations is a design
True
N
notable
constellations
design
A notable constellations is a notable design
False
N
notable
constellations
design
A possible constellations is a constellations
False
O
possible
constellations
design
A possible constellations is a design
False
O
possible
constellations
design
A possible constellations is a possible design
True
O
possible
constellations
design
A proposed constellations is a constellations
False
O
proposed
constellations
design
A proposed constellations is a design
False
O
proposed
constellations
design
A proposed constellations is a proposed design
True
O
proposed
constellations
design
A future constellations is a constellations
False
O
future
constellations
design
A future constellations is a design
False
O
future
constellations
design
A future constellations is a future design
True
O
future
constellations
design
A former constellations is a constellations
False
O
former
constellations
design
A former constellations is a design
False
O
former
constellations
design
A former constellations is a former design
True
O
former
constellations
design
A wooden contrast is a contrast
True
I
wooden
contrast
opposition
A wooden contrast is a opposition
True
I
wooden
contrast
opposition
A wooden contrast is a wooden opposition
True
I
wooden
contrast
opposition
A basic contrast is a contrast
True
N
basic
contrast
opposition
A basic contrast is a opposition
True
N
basic
contrast
opposition
A basic contrast is a basic opposition
False
N
basic
contrast
opposition
A former contrast is a contrast
False
O
former
contrast
opposition
A former contrast is a opposition
False
O
former
contrast
opposition
A former contrast is a former opposition
True
O
former
contrast
opposition
A possible contrast is a contrast
False
O
possible
contrast
opposition
A possible contrast is a opposition
False
O
possible
contrast
opposition
A possible contrast is a possible opposition
True
O
possible
contrast
opposition
A apparent contrast is a contrast
False
O
apparent
contrast
opposition
A apparent contrast is a opposition
False
O
apparent
contrast
opposition
A apparent contrast is a apparent opposition
True
O
apparent
contrast
opposition
A potential contrast is a contrast
False
O
potential
contrast
opposition
A potential contrast is a opposition
False
O
potential
contrast
opposition
A potential contrast is a potential opposition
True
O
potential
contrast
opposition
A annual revenue is a revenue
True
I
annual
revenue
sum
A annual revenue is a sum
True
I
annual
revenue
sum
A annual revenue is a annual sum
True
I
annual
revenue
sum
A total revenue is a revenue
True
N
total
revenue
sum
A total revenue is a sum
True
N
total
revenue
sum
A total revenue is a total sum
False
N
total
revenue
sum
A former revenue is a revenue
False
O
former
revenue
sum
A former revenue is a sum
False
O
former
revenue
sum
A former revenue is a former sum
True
O
former
revenue
sum
A future revenue is a revenue
False
O
future
revenue
sum
A future revenue is a sum
False
O
future
revenue
sum
A future revenue is a future sum
True
O
future
revenue
sum
A predicted revenue is a revenue
False
O
predicted
revenue
sum
A predicted revenue is a sum
False
O
predicted
revenue
sum
A predicted revenue is a predicted sum
True
O
predicted
revenue
sum
A past revenue is a revenue
False
O
past
revenue
sum
A past revenue is a sum
False
O
past
revenue
sum
A past revenue is a past sum
True
O
past
revenue
sum
A portuguese top is a top
True
I
portuguese
top
region
A portuguese top is a region
True
I
portuguese
top
region
A portuguese top is a portuguese region
True
I
portuguese
top
region
A right top is a top
True
N
right
top
region
A right top is a region
True
N
right
top
region
A right top is a right region
False
N
right
top
region
A mock top is a top
False
O
mock
top
region
A mock top is a region
False
O
mock
top
region
A mock top is a mock region
True
O
mock
top
region
A future top is a top
False
O
future
top
region

Preprocessed from https://huggingface.co/datasets/lorenzoscottb/PLANE-ood/

df=pd.read_json('https://huggingface.co/datasets/lorenzoscottb/PLANE-ood/resolve/main/PLANE_trntst-OoV_inftype-all.json')
f = lambda df: pd.DataFrame(list(zip(*[df[c] for c in df.index])),columns=df.index)
ds=DatasetDict()
for split in ['train','test']:
    dfs=pd.concat([f(df[c]) for c in df.columns if split in c.lower()]).reset_index(drop=True)
    dfs['label']=dfs['label'].map(lambda x:{1:'entailment',0:'not-entailment'}[x])
    ds[split]=Dataset.from_pandas(dfs,preserve_index=False)
ds.push_to_hub('tasksource/PLANE-ood')

PLANE Out-of-Distribution Sets

PLANE (phrase-level adjective-noun entailment) is a benchmark to test models on fine-grained compositional inference. The current dataset contains five sampled splits, used in the supervised experiments of Bertolini et al., 22.

Features

Each entrance has 6 features: seq, label, Adj_Class, Adj, Nn, Hy

  • seq:test sequense
  • label: ground truth (1:entialment, 0:no-entailment)
  • Adj_Class: the class of the sequence adjectives
  • Adj: the adjective of the sequence (I: intersective, S: subsective, O: intensional)
  • Nn: the noun
  • Hy: the noun's hypericum

Each sample in seq can take one of three forms (or inference types, in paper):

  • An Adjective-Noun is a Noun (e.g. A red car is a car)
  • An Adjective-Noun is a Hypernym(Noun) (e.g. A red car is a vehicle)
  • An Adjective-Noun is a Adjective-Hypernym(Noun) (e.g. A red car is a red vehicle)

Please note that, as specified in the paper, the ground truth is automatically assigned based on the linguistic rule that governs the interaction between each adjective class and inference type – see the paper for more detail.

Cite

If you use PLANE for your work, please cite the main COLING 2022 paper.

@inproceedings{bertolini-etal-2022-testing,
    title = "Testing Large Language Models on Compositionality and Inference with Phrase-Level Adjective-Noun Entailment",
    author = "Bertolini, Lorenzo  and
      Weeds, Julie  and
      Weir, David",
    booktitle = "Proceedings of the 29th International Conference on Computational Linguistics",
    month = oct,
    year = "2022",
    address = "Gyeongju, Republic of Korea",
    publisher = "International Committee on Computational Linguistics",
    url = "https://aclanthology.org/2022.coling-1.359",
    pages = "4084--4100",
}
Downloads last month
36
Edit dataset card