Dataset Viewer
Auto-converted to Parquet Duplicate
query_id
stringlengths
12
34
sense_id
stringlengths
9
30
lexeme_id
stringlengths
2
15
headword
stringlengths
2
15
pos
stringclasses
9 values
sense_index
int32
0
3
domain
stringclasses
160 values
tier
stringclasses
2 values
style
stringclasses
8 values
text
stringlengths
2
161
headword_free
bool
2 classes
aaa:noun:0#q0
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
keyword
car breakdown roadside help membership
true
aaa:noun:0#q1
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
question
Who can help if my car breaks down on the road?
true
aaa:noun:0#q2
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
conversational
My car won't start and I need someone to come out, who should I call?
true
aaa:noun:0#q3
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
constraint
roadside assistance without buying a full travel package
true
aaa:noun:0#q4
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
role
for a new driver needing emergency vehicle help
true
aaa:noun:0#q5
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
example_based
give me a real example of using AAA after a breakdown
false
aaa:noun:0#q6
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
step_by_step
How do I request roadside assistance step by step?
true
aaa:noun:0#q7
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
directive
compare roadside assistance with towing insurance
true
aaa:noun:0#q8
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
keyword
organization that helps stranded motorists
true
aaa:noun:0#q9
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
question
Is there a membership service that can tow my disabled car?
true
aaa:noun:0#q10
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
conversational
I need travel planning and roadside help from the same automobile club
true
aaa:noun:0#q11
aaa:noun:0
aaa
aaa
noun
0
everyday_life.transportation
core
directive
explain what AAA membership covers when a car won't start
false
aaa:noun:1#q0
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
keyword
highest credit rating for bonds
true
aaa:noun:1#q1
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
question
What rating shows that an issuer is considered exceptionally unlikely to default?
true
aaa:noun:1#q2
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
conversational
I'm looking at a company's bonds and need to know what the top possible credit grade means
true
aaa:noun:1#q3
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
constraint
creditworthiness rating for debt with minimal default risk
true
aaa:noun:1#q4
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
role
explain this to a first-time bond investor
true
aaa:noun:1#q5
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
example_based
Give me a sentence using AAA for a bond issuer
false
aaa:noun:1#q6
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
step_by_step
How do rating agencies assess whether a bond deserves the top grade?
true
aaa:noun:1#q7
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
directive
compare the highest debt rating with investment-grade ratings below it
true
aaa:noun:1#q8
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
keyword
top agency grade for corporate debt
true
aaa:noun:1#q9
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
question
Would a sovereign bond with the strongest possible repayment outlook receive this rating?
true
aaa:noun:1#q10
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
example_based
Show me a real example of an issuer whose bonds earned the highest credit grade
true
aaa:noun:1#q11
aaa:noun:1
aaa
aaa
noun
1
business.finance
core
constraint
AAA rating after financial reforms, not the roadside assistance organization
false
aaa:noun:2#q0
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
keyword
small cylinder battery for remote controls
true
aaa:noun:2#q1
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
question
Which battery size do most TV remotes use?
true
aaa:noun:2#q2
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
conversational
My toy needs the skinny little batteries, but I can't remember what size they are
true
aaa:noun:2#q3
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
constraint
remote-control battery that is small and cylindrical, not a coin cell
true
aaa:noun:2#q4
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
role
for a parent replacing batteries in a child's toy, which size should I buy?
true
aaa:noun:2#q5
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
example_based
Show me a sentence using AAA for the battery size, not the automobile association
false
aaa:noun:2#q6
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
step_by_step
How do I replace the batteries in a remote control step by step?
true
aaa:noun:2#q7
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
directive
List the portable household devices that commonly use AAA batteries
false
aaa:noun:2#q8
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
keyword
thin dry-cell batteries for toy cars
true
aaa:noun:2#q9
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
question
Do wireless mice and small toys take this kind of cylindrical battery?
true
aaa:noun:2#q10
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
conversational
I found AAA printed inside my remote and need to know which batteries to purchase
false
aaa:noun:2#q11
aaa:noun:2
aaa
aaa
noun
2
technology.hardware_devices
core
directive
Compare AAA with AA batteries for powering a handheld remote
false
aaa:noun:3#q0
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
keyword
identity access accounting security framework
true
aaa:noun:3#q1
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
question
How do network devices verify users and control what they can access?
true
aaa:noun:3#q2
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
conversational
I'm setting up centralized login controls for routers and need to track who used what
true
aaa:noun:3#q3
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
constraint
secure access management for multiple devices without separate local accounts
true
aaa:noun:3#q4
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
role
for a network administrator explaining centralized authentication to a new team member
true
aaa:noun:3#q5
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
example_based
show a realistic example of authentication, authorization, and usage tracking on a network
true
aaa:noun:3#q6
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
step_by_step
walk me through how a device checks identity, grants permissions, and records activity
true
aaa:noun:3#q7
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
directive
explain the three parts of AAA in network security
false
aaa:noun:3#q8
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
keyword
AAA meaning in network device security not roadside assistance
false
aaa:noun:3#q9
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
question
Which framework combines login verification, permission checks, and accounting logs?
true
aaa:noun:3#q10
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
conversational
I need the term for managing network users centrally and recording their activity
true
aaa:noun:3#q11
aaa:noun:3
aaa
aaa
noun
3
technology.information_security
core
directive
compare AAA with keeping separate usernames and passwords on every router
false
aaa:interjection:0#q0
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
keyword
reflex cry from sudden pain
true
aaa:interjection:0#q1
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
question
What sound might someone make immediately after being pricked by a needle?
true
aaa:interjection:0#q2
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
conversational
I need the little cry people make when an injury suddenly hurts
true
aaa:interjection:0#q3
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
constraint
a reaction to pain, not surprise or shock
true
aaa:interjection:0#q4
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
role
for a first-aid trainee, what brief vocal reaction signals sudden discomfort?
true
aaa:interjection:0#q5
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
example_based
Give an example sentence where someone says aaa after getting hurt
false
aaa:interjection:0#q6
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
step_by_step
walk me through how a person might react vocally after a painful puncture
true
aaa:interjection:0#q7
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
directive
distinguish the pain reaction from a cry of surprise
true
aaa:interjection:0#q8
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
keyword
short exclamation after touching something painfully hot
true
aaa:interjection:0#q9
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
question
Would someone use this cry when a wound suddenly starts hurting?
true
aaa:interjection:0#q10
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
conversational
I’m trying to remember the interjection that slips out when pain catches you off guard
true
aaa:interjection:0#q11
aaa:interjection:0
aaa
aaa
interjection
0
health.first_aid
core
example_based
sentence using aaa for immediate physical discomfort
false
aaa:interjection:1#q0
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
keyword
sudden surprise exclamation
true
aaa:interjection:1#q1
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
question
What do people shout when something unexpected happens?
true
aaa:interjection:1#q2
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
conversational
The lights suddenly went out and I need the right startled reaction
true
aaa:interjection:1#q3
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
constraint
an exclamation for surprise, not pain
true
aaa:interjection:1#q4
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
role
for a teacher explaining startled reactions to children
true
aaa:interjection:1#q5
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
example_based
Give a sentence using aaa for sudden alarm
false
aaa:interjection:1#q6
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
step_by_step
walk me through what someone might exclaim when a door suddenly slams
true
aaa:interjection:1#q7
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
directive
explain the difference between a startled cry and a pain reaction
true
aaa:interjection:1#q8
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
keyword
unexpected event alarm exclamation
true
aaa:interjection:1#q9
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
question
Which short cry fits the moment you suddenly realize something is wrong?
true
aaa:interjection:1#q10
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
conversational
I heard a crash upstairs and blurted something out—what kind of interjection is that?
true
aaa:interjection:1#q11
aaa:interjection:1
aaa
aaa
interjection
1
people_society.emotion_attitude
core
constraint
find a brief reaction for shock or realization, with no injury involved
true
aachen:noun:0#q0
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
keyword
German city near Belgium and the Netherlands
true
aachen:noun:0#q1
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
question
Which German city is close to both Belgium and the Netherlands?
true
aachen:noun:0#q2
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
conversational
I'm planning a trip to the border area and want a historic German city with a major university
true
aachen:noun:0#q3
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
constraint
find a university town in western Germany, not Berlin or Munich
true
aachen:noun:0#q4
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
role
for a student choosing a German university town near two international borders
true
aachen:noun:0#q5
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
example_based
Give me an example of a sentence about Aachen hosting RWTH
false
aachen:noun:0#q6
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
step_by_step
how do I plan a day exploring medieval sites and the university district there
true
aachen:noun:0#q7
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
directive
compare this border city with nearby Belgian and Dutch towns
true
aachen:noun:0#q8
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
keyword
medieval heritage western Germany university town
true
aachen:noun:0#q9
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
question
Is there a German city where Belgium and the Netherlands are both nearby?
true
aachen:noun:0#q10
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
conversational
What does Aachen actually look like as a place to visit beyond its university?
false
aachen:noun:0#q11
aachen:noun:0
aachen
aachen
noun
0
people_society.general
tier2
directive
list historic attractions in Aachen for a short visit
false
aardvark:noun:0#q0
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
keyword
African night mammal that eats ants and termites
true
aardvark:noun:0#q1
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
question
Which animal digs for termites after dark in the African savanna?
true
aardvark:noun:0#q2
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
conversational
I saw a strange long-eared animal rooting around at night—what could it be?
true
aardvark:noun:0#q3
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
constraint
identify an African termite eater without using bats or anteaters
true
aardvark:noun:0#q4
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
role
explain it to a child learning about African wildlife
true
aardvark:noun:0#q5
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
example_based
sentence using aardvark in a wildlife report
false
aardvark:noun:0#q6
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
step_by_step
how does this animal find and dig out termites step by step
true
aardvark:noun:0#q7
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
directive
describe the mammal with a pig-like nose and powerful digging claws
true
aardvark:noun:0#q8
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
keyword
nocturnal savanna animal long ears tubular teeth
true
aardvark:noun:0#q9
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
question
What animal has a piglike snout but is unrelated to pigs and feeds on ants?
true
aardvark:noun:0#q10
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
question
Are aardvarks active during the day or mainly at night?
false
aardvark:noun:0#q11
aardvark:noun:0
aardvark
aardvark
noun
0
nature.animals
tier2
example_based
show me what an African burrowing insect-eater looks like
true
aaron:noun:0#q0
aaron:noun:0
aaron
aaron
noun
0
people_society.personal_names
core
keyword
English male first name Hebrew origin
true
aaron:noun:0#q1
aaron:noun:0
aaron
aaron
noun
0
people_society.personal_names
core
question
Which Hebrew-origin boys' names are commonly used in English?
true
aaron:noun:0#q2
aaron:noun:0
aaron
aaron
noun
0
people_society.personal_names
core
conversational
I'm looking for the name people spell Aaron, not information about the biblical priest.
false
aaron:noun:0#q3
aaron:noun:0
aaron
aaron
noun
0
people_society.personal_names
core
constraint
male given name with Hebrew roots, not a religious biography
true
End of preview. Expand in Data Studio

Superseded by OpenGloss v2.1 (2026-09-07): 109,633 lexemes and 250,003 live senses — twice this release's coverage — plus a new opengloss-v2.1-inflections form→lemma lookup. v2.0 stays published for reproducibility.

OpenGloss v2.0 — Queries

Synthetic search queries written per sense, in eight styles — keyword, question, conversational, constraint, role, example-based, step-by-step and directive — with each sense's sibling senses in the prompt so the queries discriminate between the meanings of one headword. headword_free marks the queries that never name their own headword, the property that stops a retriever trained on them from collapsing into string matching. This is the doc2query side of the release; the graded relevance judgements live in opengloss-v2.0-qrels.

Part of the OpenGloss v2.0 release family — 15 datasets built from one store of 54,724 lexemes and 137,314 live senses, all joinable on derived ids. See Related datasets for the rest.

What's new in v2.0 vs v1.3

  1. Schema v3. Every lexeme carries a kind discriminator (simplex, compound, phrasal verb, idiom, proper noun, abbreviation, affix, function word); every sense carries a controlled domain leaf from a fixed ~160-leaf taxonomy instead of free text; every example carries the character span of the headword occurrence inside it.
  2. Renditions, not one string. A definition is a set: the canonical one plus rewrites at four reading levels and in four registers, each produced in a single call from the canonical text so they say the same thing at different altitudes.
  3. A sense graph, not a word graph. Typed relations resolve to sense ids wherever the target's entry exists in the release, so bank --hypernym--> financial institution points at a meaning rather than at a string.
  4. Retrieval data is first-class. Synthetic per-sense queries in eight styles, grounded QA pairs, mined word-in-context pairs, MS MARCO-style triples with graph-derived hard negatives, and graded TREC qrels — all derivable from, and consistent with, the same entries.
  5. Derivable identifiers everywhere. v1.3 published a positional id for lexemes and senses (3d_model_noun_0) and nothing below that. v2.0 gives every rendition, edge, query, QA pair and provenance record an id computable from the row alone, and never renumbers: a retired sense is tombstoned, so the ids after it keep their meaning.
  6. Per-field provenance. Which model wrote a field, how many tokens it took, what it cost — published as its own dataset.

Scope: fewer headwords, far more per headword

v2.0 is not a superset of v1.3. It covers 54,724 lexemes — a frequency-ranked subset of v1.3's 205,983 — and spends the difference on depth. If you need breadth of vocabulary, use v1.3; if you need graded renditions, resolved relations, spans, or retrieval supervision, use v2.0.

v1.3 v2.0
Lexemes 205,983 54,724
Senses 565,604 137,314
Definition renditions per sense 1 canonical 1 canonical + up to 8 graded
Relation targets bare strings resolved to sense ids
Retrieval training data companion sets queries, QA, triples, qrels
Per-field provenance no model, tokens and cost per call

Key statistics

Lexemes 54,724
Live senses 137,314
Rows in this dataset 1,330,311
Queries 1,330,311
Headword-free 1,038,924 (78.1%)
Query styles 8

By tier

Tier Lexemes Live senses
core 10,000 34,015
tier2 31,886 76,855
tier3 12,838 26,444

Coverage by tier

The release was built in three frequency-ranked passes and they did not all receive the same stages. This table is per-field and per-tier so the gaps are visible rather than averaged away.

Field Of core tier2 tier3
Canonical gloss sense 100.0% 100.0% 100.0%
Controlled domain tag sense 100.0% 100.0% 100.0%
Gloss at 4 reading levels sense 100.0% 99.9% 99.9%
Gloss in 4 registers sense 100.0% 100.0% 0.0%
At least one example sense 100.0% 99.9% 99.8%
Examples at 4 reading levels sense 99.0% 99.6% 99.8%
At least one relation sense 96.8% 97.3% 98.0%
Synthetic retrieval queries sense 100.0% 100.0% 0.0%
Grounded QA pairs sense 99.8% 99.6% 0.0%
Etymology lexeme 100.0% 100.0% 99.8%
Lexical explanation lexeme 100.0% 100.0% 100.0%
Encyclopedia (neutral) lexeme 100.0% 100.0% 100.0%
Encyclopedia at grade 5 + college (core entries also carry grade 1 and grade 10) lexeme 100.0% 100.0% 100.0%
Contrast paragraphs lexeme 72.8% 55.1% 0.0%

Query styles

Style Queries
question 229,538
keyword 220,127
conversational 180,903
example_based 172,231
directive 169,545
constraint 132,513
step_by_step 112,977
role 112,477

Files

Files Config Rows Shards Size
data/train-*.parquet default 1,330,311 3 43.3 MB

Fields

1,330,311 rows, one row per synthetic query.

Field Type Description
query_id string Derived, positional, zero-based: {sense_id}#q{n}. The stage only appends, so an id never changes meaning.
sense_id string Sense id: {lexeme_id}:{pos}:{index}. Join key.
lexeme_id string Owning entry id: slugify(headword). Join key.
headword string The owning entry's surface headword.
pos string Part of speech of the owning POS entry (noun, verb, …).
sense_index int32 Zero-based position of the sense within its POS entry.
domain string Controlled domain leaf, root.leaf (nullable).
tier string core, tier2 or tier3.
style string keyword, question, conversational, constraint, role, example_based, step_by_step or directive.
text string The query, 1–200 characters.
headword_free bool True when the query contains no form of its own headword (whole-word, case-insensitive, inflections included).

One real row:

{
  "query_id": "aaa:noun:0#q0",
  "sense_id": "aaa:noun:0",
  "lexeme_id": "aaa",
  "headword": "aaa",
  "pos": "noun",
  "sense_index": 0,
  "domain": "everyday_life.transportation",
  "tier": "core",
  "style": "keyword",
  "text": "car breakdown roadside help membership",
  "headword_free": true
}

Loading it

from datasets import load_dataset

ds = load_dataset("mjbommar/opengloss-v2.0-queries", split="train")
print(ds)
print(ds[0])

The shards are plain parquet, so nothing forces you through datasets — read them straight, locally or over hf://:

import polars as pl

df = pl.read_parquet("hf://datasets/mjbommar/opengloss-v2.0-queries/data/train-*.parquet")
print(df.head())
import duckdb

duckdb.sql("SELECT count(*) FROM 'hf://datasets/mjbommar/opengloss-v2.0-queries/data/train-*.parquet'").show()

Query/definition pairs for a bi-encoder

import polars as pl

queries = pl.read_parquet("hf://datasets/mjbommar/opengloss-v2.0-queries/data/train-*.parquet")
defs = pl.read_parquet(
    "hf://datasets/mjbommar/opengloss-v2.0-definitions/data/train-*.parquet"
).filter(pl.col("is_canonical"))
pairs = queries.join(defs.select("sense_id", "text"), on="sense_id", suffix="_gloss")
print(pairs.select("text", "text_gloss").head())

Identifiers, and how they compose

Every id is derived from structure, never randomly minted, so a consumer can recompute one from a row and join across the whole family without a lookup table. Sense positions are stable across regenerations: a retired sense is tombstoned, not removed, so the indices after it never shift.

Id Shape Example
Lexeme slugify(headword) abseil
Sense {lexeme_id}:{pos}:{index} (zero-based) abseil:verb:0
Rendition {owner_id}#{reading_level}/{register} abseil:verb:0#grade_5/plain
Entry-level owner {lexeme_id}:encyclopedia / :explanation abseil:encyclopedia
Edge {source_sense_id}-{type}->{target_lexeme_id} abseil:verb:0-synonym->rappel
Query {sense_id}#q{n} (zero-based) abseil:verb:0#q3
QA pair {sense_id}#qa{n} (zero-based) abseil:verb:0#qa3
Provenance record p{n} within its entry (one-based) p12

An edge id keys on the target's slug, not on the target's sense, so resolving a target never changes the id of the edge that found it.

Reading levels and registers

A rendition is keyed on a (reading_level, register) pair. The canonical rendition of every field is (neutral, plain); everything else is a rewrite of it.

reading_level Who it is written for Rough CCSS band
neutral The canonical text: an adult general reader, no level targeted —
grade_1 Beginning readers; short sentences, common words K–1
grade_5 Upper elementary 4–5
grade_10 Secondary 9–10
college Undergraduate and above; technical vocabulary allowed 11–CCR
register What changes Reading it
plain Nothing — the neutral register The default
informal Conversational, contractions, everyday words How you'd say it to a friend
formal Full forms, precise hedging, no contractions How you'd write it in a report
technical Domain vocabulary, exact conditions How a specialist would state it
marketing Benefit-first, persuasive framing A genre, not a formality level

marketing sits on the register axis for convenience but is a genre value rather than a point on the formality scale — worth remembering if you train a formality classifier on this column.

Related datasets

Everything below is built from the same store and joins on lexeme_id / sense_id.

Dataset Grain What it holds
opengloss-v2.0-lexicon one row per lexeme One row per lexeme: kind, morphology, etymology, encyclopedia, contrasts, sense ids, provenance summary.
opengloss-v2.0-senses one row per live sense One row per live sense: canonical gloss, 8 gloss renditions, examples, resolved relations, synthetic queries, grounded QA pairs.
opengloss-v2.0-definitions one row per gloss rendition One row per gloss rendition (canonical included): reading level, register, text, readability grade.
opengloss-v2.0-examples one row per example rendition One row per example sentence with the headword's character span, its reading level and register.
opengloss-v2.0-encyclopedia one row per encyclopedia rendition · one row per lexical-explanation rendition One row per encyclopedia article rendition, plus an explanation config for the "why this word" prose.
opengloss-v2.0-etymology one row per entry with an etymology One row per entry with an etymology: prose summary, ordered language trail, cognates, references.
opengloss-v2.0-relations one row per live relation edge · one row per removed relation edge One row per semantic edge, resolved to target sense ids; a tombstoned config recovers the edges the reconcile pass removed.
opengloss-v2.0-queries (this one) one row per synthetic query One row per synthetic retrieval query, across eight query styles, tagged to the sense it should retrieve.
opengloss-v2.0-qa-pairs one row per question/answer pair One row per grounded question/answer pair, with the rendition ids the answer cites.
opengloss-v2.0-contrasts one row per contrast paragraph One row per "X vs Y" paragraph on a synonym/antonym/confusable edge, with a verdict on the edge.
opengloss-v2.0-provenance one row per provenance record One row per recorded generation call: stage, model, tokens, cost, run id — the audit trail.
opengloss-v2.0-retrieval-pairs one row per mined pair Word-in-context and doc2query-shaped (text_a, text_b, label) pairs mined from the store for free.
opengloss-v2.0-retrieval-triples one row per (query, positive, negative) triple MS MARCO-style (query, positive, negative) triples whose hard negatives come from the graph.
opengloss-v2.0-qrels one row per query, with its whole graded candidate list · one row per document in the retrieval corpus Graded TREC relevance judgements (0–3) plus the document corpus and listwise candidate lists.
opengloss-v2.0-pretrain one row per rendered document Entries serialised into plain-prose dictionary, thesaurus, encyclopedia and usage-note documents.

Known limitations

  • It is synthetic. Every string here was written by a language model against a schema, not transcribed from a corpus or checked by a lexicographer. It is well-formed and internally consistent; it is not attested usage, and it will contain confident errors. Do not use it as ground truth about what a word means.
  • Judge scores 70.2/100 (core + tier 2) and 66.7/100 (tier 3). A different model family (Claude Opus) scored fixed 40-entry stratified samples at the close of each build. Sample statistics, not per-entry guarantees, and the judge is itself a model.
  • Relation precision is the weakest axis. Relations were judged for validity and the ones that failed were demoted rather than asserted; symmetric reciprocity finished at 98.0% for synonyms and 99.1% for antonyms, and 3,709 senses were left with no relation at all. Treat a single edge as a hypothesis, not a fact; treat the aggregate graph as usable.
  • Tier 3 is deliberately partial. 12,838 lexemes received the text stages (glosses, examples, encyclopedia) but not the queries, QA pairs, contrasts or register renditions. The coverage table above gives the exact per-field share; nothing is hidden behind an average.
  • The encyclopedia is entry-level. One article per headword, about the headword as a whole. On a polysemous entry it is not a description of any one sense, and it is never used as a positive for one (D-71). It is entry-level reference prose, not a specialist article.
  • This repo is core + tier 2 only. The tier-3 slice (12,838 lexemes) never ran this stage, so its senses are absent here entirely rather than present-and-empty. Join against opengloss-v2.0-senses if you need to know which senses have nothing.

Citation

@misc{bommarito2025opengloss,
  title  = {OpenGloss: A Synthetic Encyclopedic Dictionary and Semantic Knowledge Graph},
  author = {Bommarito, Michael J., II},
  year   = {2025},
  eprint = {2511.18622},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2511.18622}
}

License

Released under Creative Commons Attribution 4.0 International (CC-BY 4.0). Attribution to the OpenGloss project is required; commercial use is permitted.

Downloads last month
40

Paper for mjbommar/opengloss-v2.0-queries