input stringlengths 24 418 | output unknown | schema unknown |
|---|---|---|
Le Guin influenced many other authors , including Booker Prize winner Salman Rushdie , David Mitchell , Neil Gaiman , and Iain Banks . | {
"entities": {
"author": [
{
"text": "Le Guin",
"start": 0,
"end": 7
},
{
"text": "Salman Rushdie",
"start": 70,
"end": 84
},
{
"text": "David Mitchell",
"start": 87,
"end": 101
},
{
"text": ... | {
"entities": [
{
"label": "politician",
"description": "An entity of type politician, in the Real Person group of Person entities; DynamicNER category Person > Real Person > Politician.",
"description_source": "mneb"
},
{
"label": "artist",
"description": "An entity of type ... |
Popular loss functions include the hinge loss ( for linear SVMs ) and the log | {
"entities": {
"metrics": [
{
"text": "loss functions",
"start": 8,
"end": 22
},
{
"text": "hinge loss",
"start": 35,
"end": 45
},
{
"text": "log",
"start": 74,
"end": 77
}
],
"algorithm": [
... | {
"entities": [
{
"label": "politician",
"description": "An entity of type politician, in the Real Person group of Person entities; DynamicNER category Person > Real Person > Politician.",
"description_source": "mneb"
},
{
"label": "artist",
"description": "An entity of type ... |
In the "Bird's Nest" of the Beijing National Stadium, a grand track and field competition is being held, and the audience is enthusiastic and cheering for the athletes. | {
"entities": {
"sports facility": [
{
"text": "Bird's Nest",
"start": 8,
"end": 19
},
{
"text": "Beijing National Stadium",
"start": 28,
"end": 52
}
],
"sporting event": [
{
"text": "track and field competition",
... | {
"entities": [
{
"label": "politician",
"description": "An entity of type politician, in the Real Person group of Person entities; DynamicNER category Person > Real Person > Politician.",
"description_source": "mneb"
},
{
"label": "artist",
"description": "An entity of type ... |
"The Godfather," directed by Francis Ford Coppola, is widely regarded as one of the greatest crime films in cinematic history. | {
"entities": {
"film": [
{
"text": "The Godfather",
"start": 1,
"end": 14
}
],
"director": [
{
"text": "Francis Ford Coppola",
"start": 29,
"end": 49
}
]
}
} | {
"entities": [
{
"label": "politician",
"description": "An entity of type politician, in the Real Person group of Person entities; DynamicNER category Person > Real Person > Politician.",
"description_source": "mneb"
},
{
"label": "artist",
"description": "An entity of type ... |
He then pursued advanced studies at the Illinois Institute of Technology , California School of Arts and Crafts , the Academie de la Grande Chaumiere and the University of Miami . | {
"entities": {
"school": [
{
"text": "Illinois Institute of Technology",
"start": 40,
"end": 72
},
{
"text": "California School of Arts and Crafts",
"start": 75,
"end": 111
},
{
"text": "Academie de la Grande Chaumiere",
... | {
"entities": [
{
"label": "politician",
"description": "An entity of type politician, in the Real Person group of Person entities; DynamicNER category Person > Real Person > Politician.",
"description_source": "mneb"
},
{
"label": "artist",
"description": "An entity of type ... |
"The miasma theory once suggested that diseases were caused by \"bad air\" from decomposing matter, (...TRUNCATED) | {"entities":{"medical theory":[{"text":"miasma theory","start":4,"end":17},{"text":"germ theory","st(...TRUNCATED) | {"entities":[{"label":"politician","description":"An entity of type politician, in the Real Person g(...TRUNCATED) |
"Hypertension, or high blood pressure, is a common disease that increases the risk of heart disease,(...TRUNCATED) | {"entities":{"disease":[{"text":"Hypertension","start":0,"end":12},{"text":"high blood pressure","st(...TRUNCATED) | {"entities":[{"label":"politician","description":"An entity of type politician, in the Real Person g(...TRUNCATED) |
"In the luxurious suite of The Ritz Carlton Hotel in Beijing, a business traveler is enjoying the be(...TRUNCATED) | {"entities":{"hotel":[{"text":"The Ritz Carlton Hotel","start":26,"end":48}],"city":[{"text":"Beijin(...TRUNCATED) | {"entities":[{"label":"politician","description":"An entity of type politician, in the Real Person g(...TRUNCATED) |
"Susan Look Avery hosted Lucy Stone and Henry Browne Blackwell of the American Woman Suffrage Associ(...TRUNCATED) | {"entities":{"actor":[{"text":"Susan Look Avery","start":0,"end":16}],"politician":[{"text":"Lucy St(...TRUNCATED) | {"entities":[{"label":"politician","description":"An entity of type politician, in the Real Person g(...TRUNCATED) |
"In the open-pit coal mine in Datong, Shanxi, huge excavators are mining coal, and transport vehicle(...TRUNCATED) | {"entities":{"mine":[{"text":"open-pit coal mine","start":7,"end":25}],"city":[{"text":"Datong","sta(...TRUNCATED) | {"entities":[{"label":"politician","description":"An entity of type politician, in the Real Person g(...TRUNCATED) |
DynamicNER (English)
DynamicNER (English) (named entity recognition (fine-grained)), rebuilt for MNEB from the original release with character-level offsets.
Source: https://github.com/Astarojth/DynamicNER (commit 81813ba8, DynamicNER.7z, base/en/BASE)
| config | split | records |
|---|---|---|
| default | train | 900 |
| default | validation | 300 |
| default | test | 300 |
Labels
entities: politician, artist, author, athlete, director, actor, scholar, military person, musician, business executive, other person, mythological figure, other fictional figure, water body, mountain, island, desert, other geographical entity, continent, country, state or province, city, district, region, other geo-political entity, address, road, railway, other address, beverages, packaged foods, other food, firearms, biological weapon, chemical weapon, explosives, cold weapon, nuclear weapon, other weapon, software, website, electronics, AI, other technology, aircraft, car, watercraft, rail vehicle, bike, other vehicle, clothes, household product, personal care product, toys, musical instruments, other product, hospital, library, park, landmark, school, museum, sports facility, other public facility, hotel, restaurant, market or mall, theater or cinema, bank, other commercial facility, airport, station, port, other transportation facility, factory, farm, mine, energy facility, other production facility, residential building, government facility, other facility, painting, sculpture, visual art genre, other visual art, song, album, music genre, other music, poem, non-fiction work, fiction work, literature genre, other literature, film, play, broadcast program, game, other art, ethnic group, religious group, other social group, educational and research organization, political or military organization, community organization, religious organization, other non-commercial organization, sports team or league, band, company, media company, other commercial organization, literary award, sports award, artistic award, other award, political or military event, sporting event, disaster, business event, other event, educational degree, tradition, god, law, language, miscellaneous, protein, species, biological theory, other biological entity, chemical element, chemical compound, chemical reaction, chemical theory, other chemical entity, physical phenomenon, astronomical object, physical theory, other physical entity, programming language, algorithm, other computer science entity, disease, injury, medication, symptom, medical theory, other medical entity, discipline, academic journal, conference, metrics, other scientific entity
Format
Each record has input (the text), output (gold annotations grouped by label) and schema (the label set with descriptions). Offsets are character offsets into input, half-open [start, end).
"output": {
"entities": {"<label>": [{"text", "start", "end"}]},
"relations": {"<label>": [{"head": {"text", "start", "end"},
"tail": {"text", "start", "end"}}]},
"events": {"<label>": [{"trigger": {"text", "start", "end"},
"arguments": [{"role", "text", "start", "end"}]}]}
}
Each schema item has a description and a description_source: official when adapted from the dataset's paper or guideline, mneb when written for MNEB. Relation datasets also give setting (end_to_end or gold_entity) and negative_label; relation items give symmetric. Event items list the roles valid for that event type.
Within a record, identical annotations are kept once and outer whitespace is trimmed from spans; source annotations are otherwise kept as annotated.
Notes
- English base version, official train/dev/test files as released (dev -> validation).
- The release gives mention strings without offsets; they are located in the sentence, and strings not found verbatim are dropped and counted.
- All 155 fine types of the hierarchy are listed; some type names are expanded for clarity, and descriptions are written for MNEB from the hierarchy.
Build
Built by scripts/datasets/dynamic_ner.py in the MNEB repository.
Citation
Luo et al. (2025), DynamicNER: A Dynamic, Multilingual, and Fine-Grained Dataset for LLM-based Named Entity Recognition. EMNLP.
License: apache-2.0
- Downloads last month
- 215