mjbommar/opengloss-v1.3-encyclopedia-variants

---
license: cc-by-4.0
task_categories:
- text-generation
- summarization
language:
- en
tags:
- style-transfer
- paraphrase
- text-simplification
- encyclopedia
- lexicon
- synthetic
- education
- opengloss
- writing-style
size_categories:
- 10K<n<100K
---

# OpenGloss Encyclopedia Variants v1.3

## Dataset Summary

**OpenGloss Encyclopedia Variants** is a synthetic dataset of vocabulary encyclopedia entries
rewritten in multiple writing styles. Each record contains an academic base entry alongside
a variant rewritten for a specific audience, tone, and content structure.

This dataset supports style transfer, text simplification, paraphrase generation, and
audience-adaptive content creation. It is derived from the
[OpenGloss](https://huggingface.co/datasets/mjbommar/opengloss-v1.3-definitions)
encyclopedic dictionary.

### Key Statistics

- **133,135 encyclopedia variants**
- **133,094 unique words/phrases**
- **133,135 unique lexemes**
- **10 writing voices** × **9 tones** × **8 framings** × **3 lengths**
- **Average base entry: 297 words** (2195 chars)
- **Average variant: 130 words** (886 chars)

### Style Dimensions

Each variant is characterized by four style dimensions:

| Dimension | Values | Description |
|-----------|--------|-------------|
| **Voice** | 10 | Writing persona (teen_explainer, tech_writer, storyteller, etc.) |
| **Tone** | 9 | Emotional register (myth_buster, narrative_hook, q_and_a, etc.) |
| **Framing** | 8 | Content structure (faq, origin_story, comparisons, etc.) |
| **Length** | 3 | Content length (short, medium, long) |

### POS Distribution

| Part of Speech | Count |
|----------------|-------|
| noun | 95,287 |
| adjective | 23,306 |
| verb | 10,862 |
| adverb | 2,959 |
| determiner | 364 |
| preposition | 166 |
| interjection | 96 |
| pronoun | 62 |
| conjunction | 30 |
| numeral | 2 |
| adjetivo | 1 |

### Voice Distribution

| Voice | Count |
|-------|-------|
| science_journalist | 133,135 |

### Tone Distribution

| Tone | Count |
|------|-------|
| plain_neutral | 133,135 |

### Framing Distribution

| Framing | Count |
|---------|-------|
| use_cases | 133,135 |

### Length Distribution

| Length | Count |
|--------|-------|
| short | 133,135 |

## Loading the Dataset

```python
from datasets import load_dataset

# Load the full dataset
dataset = load_dataset("mjbommar/opengloss-v1.3-encyclopedia-variants")

# Access records
for record in dataset["train"]:
print(f"Word: {record['word']}")
print(f"Voice: {record['voice']} | Tone: {record['tone']}")
print(f"Base: {record['base_entry'][:200]}...")
print(f"Variant: {record['variant_text'][:200]}...\n")
```

## Example Record

```python
{
"id": "orange_twist_storyteller_narrative_hook_faq_short",
"word": "orange twist",
"lexeme_id": "orange_twist",
"pos": "noun",
"base_entry": "### **Orange twist**\n\nAn **orange twist** is a slender spiral of orange peel...",
"variant_text": "What is an orange twist, really? Not just a curl of peel, but a tiny bit of theater...",
"style": {
"voice": "storyteller",
"tone": "narrative_hook",
"framing": "faq",
"length_label": "short"
},
"voice": "storyteller",
"tone": "narrative_hook",
"framing": "faq",
"length_label": "short"
}
```

## Use Cases

### Style Transfer Training
Train models to rewrite text in different styles:

```python
# Create style-conditioned pairs
for record in dataset["train"]:
source = record["base_entry"]
target = record["variant_text"]
style = f"[VOICE:{record['voice']}] [TONE:{record['tone']}]"
# Use for seq2seq training with style prefix
```

### Text Simplification
Filter for simplified variants:

```python
# Get accessible versions
simplified = dataset["train"].filter(
lambda x: x["voice"] in ["teen_explainer", "friendly_teacher"]
and x["length_label"] == "short"
)
```

### Audience-Adaptive Generation
Generate content for specific audiences:

```python
# Get policy-oriented content
policy_variants = dataset["train"].filter(
lambda x: x["voice"] == "policy_brief"
)
```

### Paraphrase Mining
Extract semantically equivalent pairs:

```python
# Same word, different styles = paraphrases
from collections import defaultdict
by_word = defaultdict(list)
for record in dataset["train"]:
by_word[record["word"]].append(record)
```

## Citation

If you use this dataset in your research, please cite:

```bibtex
@misc{bommarito2025opengloss_variants,
title={OpenGloss Encyclopedia Variants: Multi-Style Vocabulary Explanations},
author={Bommarito, Michael J., II},
year={2025},
url={https://huggingface.co/datasets/mjbommar/opengloss-v1.3-encyclopedia-variants},
note={Dataset available under CC-BY 4.0}
}
```

## License

This dataset is released under **Creative Commons Attribution 4.0 International (CC-BY 4.0)**.

## Related Datasets

- [OpenGloss v1.3 Dictionary](https://huggingface.co/datasets/mjbommar/opengloss-v1.3-dictionary) - Word-level records
- [OpenGloss v1.3 Definitions](https://huggingface.co/datasets/mjbommar/opengloss-v1.3-definitions) - Definition-level records
- [OpenGloss v1.3 Contrastive Examples](https://huggingface.co/datasets/mjbommar/opengloss-v1.3-contrastive-examples) - Semantic gradients
- [OpenGloss v1.3 Query Examples](https://huggingface.co/datasets/mjbommar/opengloss-v1.3-query-examples) - Query-side retrieval supervision
- [OpenGloss v1.3 Hard Negative Pairs](https://huggingface.co/datasets/mjbommar/opengloss-v1.3-hard-negative-pairs) - Calibration pairs for embedding training

## Acknowledgments

This dataset was generated using:
- [OpenGloss](https://huggingface.co/datasets/mjbommar/opengloss-v1.3-definitions) lexicon data
- OpenAI GPT models for style-varied generation
- [pydantic-ai](https://github.com/pydantic/pydantic-ai) for structured generation

---

*Generated from the OpenGloss v1.3 lexicon.*