README.md

---
language: srn
language_name: Sranan Tongo
language_family: germanic_west_anglofrisian
tags:
  - wikilangs
  - nlp
  - tokenizer
  - embeddings
  - n-gram
  - markov
  - wikipedia
  - feature-extraction
  - sentence-similarity
  - tokenization
  - n-grams
  - markov-chain
  - text-mining
  - fasttext
  - babelvec
  - vocabulous
  - vocabulary
  - monolingual
  - family-germanic_west_anglofrisian
license: mit
library_name: wikilangs
pipeline_tag: text-generation
datasets:
  - omarkamali/wikipedia-monthly
dataset_info:
  name: wikipedia-monthly
  description: Monthly snapshots of Wikipedia articles across 300+ languages
metrics:
  - name: best_compression_ratio
    type: compression
    value: 4.163
  - name: best_isotropy
    type: isotropy
    value: 0.1199
  - name: vocabulary_size
    type: vocab
    value: 0
generated: 2026-01-10
---

# Sranan Tongo - Wikilangs Models
## Comprehensive Research Report & Full Ablation Study

This repository contains NLP models trained and evaluated by Wikilangs, specifically on **Sranan Tongo** Wikipedia data.
We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and word embeddings.

## 📋 Repository Contents

### Models & Assets

- Tokenizers (8k, 16k, 32k, 64k)
- N-gram models (2, 3, 4, 5-gram)
- Markov chains (context of 1, 2, 3, 4 and 5)
- Subword N-gram and Markov chains
- Embeddings in various sizes and dimensions (aligned and unaligned)
- Language Vocabulary
- Language Statistics

![Performance Dashboard](visualizations/performance_dashboard.png)

### Analysis and Evaluation

- [1. Tokenizer Evaluation](#1-tokenizer-evaluation)
- [2. N-gram Model Evaluation](#2-n-gram-model-evaluation)
- [3. Markov Chain Evaluation](#3-markov-chain-evaluation)
- [4. Vocabulary Analysis](#4-vocabulary-analysis)
- [5. Word Embeddings Evaluation](#5-word-embeddings-evaluation)
- [6. Morphological Analysis (Experimental)](#6--morphological-analysis-experimental)
- [7. Summary & Recommendations](#7-summary--recommendations)
- [Metrics Glossary](#appendix-metrics-glossary--interpretation-guide)
- [Visualizations Index](#visualizations-index)

---
## 1. Tokenizer Evaluation

![Tokenizer Compression](visualizations/tokenizer_compression.png)

![Tokenizer Fertility](visualizations/tokenizer_fertility.png)

![Tokenizer OOV](visualizations/tokenizer_oov.png)

![Total Tokens](visualizations/tokenizer_total_tokens.png)

### Results

| Vocab Size | Compression | Avg Token Len | UNK Rate | Total Tokens |
|------------|-------------|---------------|----------|--------------|
| **8k** | 3.920x | 3.93 | 0.1075% | 123,705 |
| **16k** | 4.163x 🏆 | 4.17 | 0.1142% | 116,482 |

### Tokenization Examples

Below are sample sentences tokenized with each vocabulary size:

**Sample 1:** `Wan fisi e ben wan guru fu a Sabi fu libi biologisi Meti metiriki. Den fisi e li...`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁wan ▁fisi ▁e ▁ben ▁wan ▁guru ▁fu ▁a ▁sabi ▁fu ... (+23 more)` | 33 |
| 16k | `▁wan ▁fisi ▁e ▁ben ▁wan ▁guru ▁fu ▁a ▁sabi ▁fu ... (+20 more)` | 30 |

**Sample 2:** `Atlanta ben wan presi ini Kondre Makandrameki. Flaku: 343 km² Man: 420 003`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁atlan ta ▁ben ▁wan ▁presi ▁ini ▁kondre ▁makandrameki . ▁flaku ... (+17 more)` | 27 |
| 16k | `▁atlanta ▁ben ▁wan ▁presi ▁ini ▁kondre ▁makandrameki . ▁flaku : ... (+16 more)` | 26 |

**Sample 3:** `George Washington (Fostu 22 dey, – Fostwarfu 14 dey, ben wan presidenti A Kondre...`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁george ▁washington ▁( fos tu ▁ 2 2 ▁dey , ... (+14 more)` | 24 |
| 16k | `▁george ▁washington ▁( fostu ▁ 2 2 ▁dey , ▁– ... (+13 more)` | 23 |


### Key Findings

- **Best Compression:** 16k achieves 4.163x compression
- **Lowest UNK Rate:** 8k with 0.1075% unknown tokens
- **Trade-off:** Larger vocabularies improve compression but increase model size
- **Recommendation:** 32k vocabulary provides optimal balance for production use

---
## 2. N-gram Model Evaluation

![N-gram Perplexity](visualizations/ngram_perplexity.png)

![N-gram Unique](visualizations/ngram_unique.png)

![N-gram Coverage](visualizations/ngram_coverage.png)

### Results

| N-gram | Variant | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
|--------|---------|------------|---------|----------------|------------------|-------------------|
| **2-gram** | Word | 342 | 8.42 | 1,014 | 51.4% | 99.9% |
| **2-gram** | Subword | 182 🏆 | 7.51 | 1,106 | 77.6% | 99.9% |
| **3-gram** | Word | 482 | 8.91 | 1,278 | 40.5% | 98.7% |
| **3-gram** | Subword | 963 | 9.91 | 6,594 | 42.8% | 86.2% |
| **4-gram** | Word | 550 | 9.10 | 1,733 | 36.1% | 97.6% |
| **4-gram** | Subword | 2,662 | 11.38 | 22,081 | 30.3% | 70.1% |
| **5-gram** | Word | 506 | 8.98 | 1,385 | 37.3% | 98.7% |
| **5-gram** | Subword | 3,796 | 11.89 | 29,372 | 24.6% | 64.3% |

### Top 5 N-grams by Size

**2-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `e ben` | 4,382 |
| 2 | `ben wan` | 2,155 |
| 3 | `ini a` | 804 |
| 4 | `fu a` | 779 |
| 5 | `e taki` | 743 |

**3-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `e ben wan` | 1,586 |
| 2 | `yari e ben` | 695 |
| 3 | `disi e ben` | 693 |
| 4 | `a e ben` | 498 |
| 5 | `man e taki` | 489 |

**4-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `a e ben wan` | 266 |
| 2 | `yari e ben wan` | 239 |
| 3 | `e ben taki a` | 230 |
| 4 | `e ben disi e` | 229 |
| 5 | `ben disi e ben` | 229 |

**5-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `e ben disi e ben` | 229 |
| 2 | `ben leki ala yari e` | 228 |
| 3 | `e ben leki ala yari` | 228 |
| 4 | `tu no frugeti ma man` | 228 |
| 5 | `no frugeti ma man e` | 228 |

**2-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `i _` | 21,858 |
| 2 | `n _` | 16,553 |
| 3 | `a n` | 13,548 |
| 4 | `a _` | 13,330 |
| 5 | `e n` | 12,307 |

**3-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `e n _` | 8,392 |
| 2 | `_ e _` | 6,647 |
| 3 | `a n _` | 6,634 |
| 4 | `_ b e` | 6,317 |
| 5 | `b e n` | 6,136 |

**4-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `_ b e n` | 6,064 |
| 2 | `b e n _` | 4,777 |
| 3 | `e _ b e` | 4,448 |
| 4 | `_ e _ b` | 4,390 |
| 5 | `w a n _` | 4,035 |

**5-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `_ b e n _` | 4,747 |
| 2 | `e _ b e n` | 4,433 |
| 3 | `_ e _ b e` | 4,384 |
| 4 | `_ w a n _` | 3,791 |
| 5 | `_ d i s i` | 3,412 |


### Key Findings

- **Best Perplexity:** 2-gram (subword) with 182
- **Entropy Trend:** Decreases with larger n-grams (more predictable)
- **Coverage:** Top-1000 patterns cover ~64% of corpus
- **Recommendation:** 4-gram or 5-gram for best predictive performance

---
## 3. Markov Chain Evaluation

![Markov Entropy](visualizations/markov_entropy.png)

![Markov Contexts](visualizations/markov_contexts.png)

![Markov Branching](visualizations/markov_branching.png)

### Results

| Context | Variant | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
|---------|---------|-------------|------------|------------------|-----------------|----------------|
| **1** | Word | 0.4288 | 1.346 | 2.41 | 11,359 | 57.1% |
| **1** | Subword | 0.9691 | 1.958 | 6.55 | 475 | 3.1% |
| **2** | Word | 0.1256 | 1.091 | 1.25 | 27,059 | 87.4% |
| **2** | Subword | 0.8794 | 1.840 | 4.73 | 3,103 | 12.1% |
| **3** | Word | 0.0454 | 1.032 | 1.08 | 33,532 | 95.5% |
| **3** | Subword | 0.7551 | 1.688 | 3.16 | 14,656 | 24.5% |
| **4** | Word | 0.0194 🏆 | 1.014 | 1.03 | 35,649 | 98.1% |
| **4** | Subword | 0.4506 | 1.367 | 1.89 | 46,131 | 54.9% |

### Generated Text Samples (Word-based)

Below are text samples generated from each word-based Markov chain model:

**Context Size 1:**

1. `e taki dati disi e ben wili wan presi fu teri numro leki dekstiri`
2. `ben wan tipisi frugetism yari ini go ten middeleeuwen ini disi e ben nc41 burco ben`
3. `a minti arienzo e du a numro fu feti ini a mamafoto na 10 fostu instansi`

**Context Size 2:**

1. `e ben wan presi ini sranankondre stori geografi demografi legi si oktu tafra 1`
2. `ben wan man ska abra taki disi ten ini go ten abra tengi di man pramisi tu`
3. `ini a sranantongo tongo efru yu pasa no abra disi ondrowerpi yu e ben disi e ben`

**Context Size 3:**

1. `e ben wan kuri fu den owrur ten fu den medium ten middeleeuwen ini a bakratongo jesus dy`
2. `yari e ben wili wan man meki no u wi somtengi frugeti e ben disi e ben taki`
3. `disi e ben taki a salekism si oktu trawan meni fu ten tafra`

**Context Size 4:**

1. `a e ben wan ondrodeli fu a arrondissementi briey geografi a opoflaku fu aboncourt meurthe et moselle...`
2. `yari e ben wan yari nanga pasa peyna nanga ledi ma oktu nanga gu tengi disi e ben no`
3. `e ben taki a salekism si oktu tafra 82`


### Generated Text Samples (Subword-based)

Below are text samples generated from each subword-based Markov chain model:

**Context Size 1:**

1. `_n_(ma_a_ikateng`
2. `anoramalinascing`
3. `i_facoku_ten_ngu`

**Context Size 2:**

1. `i_wangi_masi_disi`
2. `n_a_otecium_re_re`
3. `an_num_res,_wariu`

**Context Size 3:**

1. `en_saleki_disi_yu_`
2. `_e_ben_on_ini_alat`
3. `an_komili_wan_the_`

**Context Size 4:**

1. `_ben_nowtu_(arabi_s`
2. `ben_wan_merkirasil_`
3. `e_ben_wan_fubenin_f`


### Key Findings

- **Best Predictability:** Context-4 (word) with 98.1% predictability
- **Branching Factor:** Decreases with context size (more deterministic)
- **Memory Trade-off:** Larger contexts require more storage (46,131 contexts)
- **Recommendation:** Context-3 or Context-4 for text generation

---
## 4. Vocabulary Analysis

![Zipf's Law](visualizations/zipf_law.png)

![Top Words](visualizations/top20_words.png)

![Coverage Curve](visualizations/vocab_coverage.png)

### Statistics

| Metric | Value |
|--------|-------|
| Vocabulary Size | 3,476 |
| Total Tokens | 93,169 |
| Mean Frequency | 26.80 |
| Median Frequency | 3 |
| Frequency Std Dev | 228.66 |

### Most Common Words

| Rank | Word | Frequency |
|------|------|-----------|
| 1 | e | 6,667 |
| 2 | ben | 6,040 |
| 3 | a | 5,047 |
| 4 | wan | 3,840 |
| 5 | fu | 3,602 |
| 6 | disi | 3,403 |
| 7 | ini | 2,307 |
| 8 | nanga | 2,251 |
| 9 | man | 1,857 |
| 10 | no | 1,719 |

### Least Common Words (from vocabulary)

| Rank | Word | Frequency |
|------|------|-----------|
| 1 | helsinki | 2 |
| 2 | voro | 2 |
| 3 | nationale | 2 |
| 4 | bedrijf | 2 |
| 5 | jari | 2 |
| 6 | winod | 2 |
| 7 | bba | 2 |
| 8 | whanau | 2 |
| 9 | pirimia | 2 |
| 10 | wahine | 2 |

### Zipf's Law Analysis

| Metric | Value |
|--------|-------|
| Zipf Coefficient | 1.1374 |
| R² (Goodness of Fit) | 0.962381 |
| Adherence Quality | **excellent** |

### Coverage Analysis

| Top N Words | Coverage |
|-------------|----------|
| Top 100 | 75.5% |
| Top 1,000 | 93.3% |
| Top 5,000 | 0.0% |
| Top 10,000 | 0.0% |

### Key Findings

- **Zipf Compliance:** R²=0.9624 indicates excellent adherence to Zipf's law
- **High Frequency Dominance:** Top 100 words cover 75.5% of corpus
- **Long Tail:** -6,524 words needed for remaining 100.0% coverage

---
## 5. Word Embeddings Evaluation

![Embedding Isotropy](visualizations/embedding_isotropy.png)

![Similarity Matrix](visualizations/embedding_similarity.png)

![t-SNE Words](visualizations/tsne_words.png)

![t-SNE Sentences](visualizations/tsne_sentences.png)


### 5.1 Cross-Lingual Alignment

![Alignment Quality](visualizations/embedding_alignment_quality.png)

![Multilingual t-SNE](visualizations/embedding_tsne_multilingual.png)


### 5.2 Model Comparison

| Model | Dimension | Isotropy | Semantic Density | Alignment R@1 | Alignment R@10 |
|-------|-----------|----------|------------------|---------------|----------------|
| **mono_32d** | 32 | 0.1199 | 0.5444 | N/A | N/A |
| **mono_64d** | 64 | 0.0180 | 0.5699 | N/A | N/A |
| **mono_128d** | 128 | 0.0021 | 0.5677 | N/A | N/A |
| **aligned_32d** | 32 | 0.1199 🏆 | 0.5345 | 0.0113 | 0.0998 |
| **aligned_64d** | 64 | 0.0180 | 0.5387 | 0.0091 | 0.1293 |
| **aligned_128d** | 128 | 0.0021 | 0.5606 | 0.0249 | 0.1406 |

### Key Findings

- **Best Isotropy:** aligned_32d with 0.1199 (more uniform distribution)
- **Semantic Density:** Average pairwise similarity of 0.5526. Lower values indicate better semantic separation.
- **Alignment Quality:** Aligned models achieve up to 2.5% R@1 in cross-lingual retrieval.
- **Recommendation:** 128d aligned for best cross-lingual performance

---
## 6.  Morphological Analysis (Experimental)

This section presents an automated morphological analysis derived from the statistical divergence between word-level and subword-level models. By analyzing where subword predictability spikes and where word-level coverage fails, we can infer linguistic structures without supervised data.

### 6.1 Productivity & Complexity

| Metric | Value | Interpretation | Recommendation |
|--------|-------|----------------|----------------|
| Productivity Index | **5.000** | High morphological productivity | Reliable analysis |
| Idiomaticity Gap | **0.537** | High formulaic/idiomatic content | - |

### 6.2 Affix Inventory (Productive Units)

These are the most productive prefixes and suffixes identified by sampling the vocabulary for global substitutability patterns. A unit is considered an affix if stripping it leaves a valid stem that appears in other contexts.

#### Productive Prefixes
| Prefix | Examples |
|--------|----------|
| `-s` | surud, smadoti, siki |
| `-a` | aban, animalia, area |
| `-b` | bisi, boosaaso, botticelli |
| `-m` | meijer, me, major |
| `-k` | kankan, kanguru, kirkedomo |
| `-p` | puspusi, peyna, part |
| `-d` | dattie, doro, damme |
| `-ma` | major, mar, mapana |

#### Productive Suffixes
| Suffix | Examples |
|--------|----------|
| `-i` | interaksi, feti, puspusi |
| `-e` | trowe, me, camille |
| `-n` | tjon, aban, granman |
| `-a` | ndyuka, tarra, animalia |
| `-s` | goolis, spijkers, ons |
| `-o` | boosaaso, kirkedomo, trio |
| `-ti` | feti, smadoti, santi |
| `-re` | bere, italiyanikondre, condre |

### 6.3 Bound Stems (Lexical Roots)

Bound stems are high-frequency subword units that are semantically cohesive but rarely appear as standalone words. These often correspond to the 'core' of a word that requires inflection or derivation to be valid.

| Stem | Cohesion | Substitutability | Examples |
|------|----------|------------------|----------|
| `anga` | 1.31x | 12 contexts | langa, nanga, ganga |
| `enti` | 1.47x | 7 contexts | efenti, sentir, peenti |

### 6.4 Affix Compatibility (Co-occurrence)

This table shows which prefixes and suffixes most frequently co-occur on the same stems, revealing the 'stacking' rules of the language's morphology.

| Prefix | Suffix | Frequency | Examples |
|--------|--------|-----------|----------|
| `-s` | `-i` | 36 words | smadoti, siki |
| `-p` | `-i` | 31 words | puspusi, prosenti |
| `-k` | `-i` | 30 words | kripi, kongri |
| `-m` | `-i` | 22 words | mindri, malsi |
| `-b` | `-i` | 20 words | bisi, botticelli |
| `-s` | `-e` | 20 words | stallone, singie |
| `-a` | `-e` | 17 words | associazione, australiyankondre |
| `-a` | `-i` | 17 words | akutimi, aktivisti |
| `-d` | `-i` | 16 words | darmi, doysri |
| `-m` | `-e` | 15 words | me, mike |

### 6.5 Recursive Morpheme Segmentation

Using **Recursive Hierarchical Substitutability**, we decompose complex words into their constituent morphemes. This approach handles nested affixes (e.g., `prefix-prefix-root-suffix`).

| Word | Suggested Split | Confidence | Stem |
|------|-----------------|------------|------|
| ukrainalanti | **`ukrainal-an-ti`** | 7.5 | `an` |
| demokrasia | **`demokra-si-a`** | 7.5 | `si` |
| politongo | **`po-li-tongo`** | 7.5 | `tongo` |
| sentralanti | **`sentral-an-ti`** | 7.5 | `an` |
| plandasie | **`planda-si-e`** | 7.5 | `si` |
| sranantaki | **`sranant-a-ki`** | 7.5 | `a` |
| importanti | **`import-an-ti`** | 7.5 | `an` |
| ondrobenin | **`ondroben-i-n`** | 6.0 | `ondroben` |
| victorien | **`victor-i-en`** | 6.0 | `victor` |
| koptisches | **`koptische-s`** | 4.5 | `koptische` |
| massimiliano | **`ma-s-similiano`** | 4.5 | `similiano` |
| nederlands | **`nederland-s`** | 4.5 | `nederland` |
| institute | **`institut-e`** | 4.5 | `institut` |
| koptische | **`koptisch-e`** | 4.5 | `koptisch` |
| verenigde | **`verenigd-e`** | 4.5 | `verenigd` |

### 6.6 Linguistic Interpretation

> **Automated Insight:**
The language Sranan Tongo shows high morphological productivity. The subword models are significantly more efficient than word models, suggesting a rich system of affixation or compounding.

> **Note on Idiomaticity:** The high Idiomaticity Gap suggests a large number of frequent multi-word expressions or formulaic sequences that are statistically distinct from their component parts.

---
## 7. Summary & Recommendations

![Performance Dashboard](visualizations/performance_dashboard.png)

### Production Recommendations

| Component | Recommended | Rationale |
|-----------|-------------|-----------|
| Tokenizer | **16k BPE** | Best compression (4.16x) |
| N-gram | **2-gram** | Lowest perplexity (182) |
| Markov | **Context-4** | Highest predictability (98.1%) |
| Embeddings | **100d** | Balanced semantic capture and isotropy |


---
## Appendix: Metrics Glossary & Interpretation Guide

This section provides definitions, intuitions, and guidance for interpreting the metrics used throughout this report.

### Tokenizer Metrics

**Compression Ratio**
> *Definition:* The ratio of characters to tokens (chars/token). Measures how efficiently the tokenizer represents text.
>
> *Intuition:* Higher compression means fewer tokens needed to represent the same text, reducing sequence lengths for downstream models. A 3x compression means ~3 characters per token on average.
>
> *What to seek:* Higher is generally better for efficiency, but extremely high compression may indicate overly aggressive merging that loses morphological information.

**Average Token Length (Fertility)**
> *Definition:* Mean number of characters per token produced by the tokenizer.
>
> *Intuition:* Reflects the granularity of tokenization. Longer tokens capture more context but may struggle with rare words; shorter tokens are more flexible but increase sequence length.
>
> *What to seek:* Balance between 2-5 characters for most languages. Arabic/morphologically-rich languages may benefit from slightly longer tokens.

**Unknown Token Rate (OOV Rate)**
> *Definition:* Percentage of tokens that map to the unknown/UNK token, indicating words the tokenizer cannot represent.
>
> *Intuition:* Lower OOV means better vocabulary coverage. High OOV indicates the tokenizer encounters many unseen character sequences.
>
> *What to seek:* Below 1% is excellent; below 5% is acceptable. BPE tokenizers typically achieve very low OOV due to subword fallback.

### N-gram Model Metrics

**Perplexity**
> *Definition:* Measures how "surprised" the model is by test data. Mathematically: 2^(cross-entropy). Lower values indicate better prediction.
>
> *Intuition:* If perplexity is 100, the model is as uncertain as if choosing uniformly among 100 options at each step. A perplexity of 10 means effectively choosing among 10 equally likely options.
>
> *What to seek:* Lower is better. Perplexity decreases with larger n-grams (more context). Values vary widely by language and corpus size.

**Entropy**
> *Definition:* Average information content (in bits) needed to encode the next token given the context. Related to perplexity: perplexity = 2^entropy.
>
> *Intuition:* High entropy means high uncertainty/randomness; low entropy means predictable patterns. Natural language typically has entropy between 1-4 bits per character.
>
> *What to seek:* Lower entropy indicates more predictable text patterns. Entropy should decrease as n-gram size increases.

**Coverage (Top-K)**
> *Definition:* Percentage of corpus occurrences explained by the top K most frequent n-grams.
>
> *Intuition:* High coverage with few patterns indicates repetitive/formulaic text; low coverage suggests diverse vocabulary usage.
>
> *What to seek:* Depends on use case. For language modeling, moderate coverage (40-60% with top-1000) is typical for natural text.

### Markov Chain Metrics

**Average Entropy**
> *Definition:* Mean entropy across all contexts, measuring average uncertainty in next-word prediction.
>
> *Intuition:* Lower entropy means the model is more confident about what comes next. Context-1 has high entropy (many possible next words); Context-4 has low entropy (few likely continuations).
>
> *What to seek:* Decreasing entropy with larger context sizes. Very low entropy (<0.1) indicates highly deterministic transitions.

**Branching Factor**
> *Definition:* Average number of unique next tokens observed for each context.
>
> *Intuition:* High branching = many possible continuations (flexible but uncertain); low branching = few options (predictable but potentially repetitive).
>
> *What to seek:* Branching factor should decrease with context size. Values near 1.0 indicate nearly deterministic chains.

**Predictability**
> *Definition:* Derived metric: (1 - normalized_entropy) × 100%. Indicates how deterministic the model's predictions are.
>
> *Intuition:* 100% predictability means the next word is always certain; 0% means completely random. Real text falls between these extremes.
>
> *What to seek:* Higher predictability for text generation quality, but too high (>98%) may produce repetitive output.

### Vocabulary & Zipf's Law Metrics

**Zipf's Coefficient**
> *Definition:* The slope of the log-log plot of word frequency vs. rank. Zipf's law predicts this should be approximately -1.
>
> *Intuition:* A coefficient near -1 indicates the corpus follows natural language patterns where a few words are very common and most words are rare.
>
> *What to seek:* Values between -0.8 and -1.2 indicate healthy natural language distribution. Deviations may suggest domain-specific or artificial text.

**R² (Coefficient of Determination)**
> *Definition:* Measures how well the linear fit explains the frequency-rank relationship. Ranges from 0 to 1.
>
> *Intuition:* R² near 1.0 means the data closely follows Zipf's law; lower values indicate deviation from expected word frequency patterns.
>
> *What to seek:* R² > 0.95 is excellent; > 0.99 indicates near-perfect Zipf adherence typical of large natural corpora.

**Vocabulary Coverage**
> *Definition:* Cumulative percentage of corpus tokens accounted for by the top N words.
>
> *Intuition:* Shows how concentrated word usage is. If top-100 words cover 50% of text, the corpus relies heavily on common words.
>
> *What to seek:* Top-100 covering 30-50% is typical. Higher coverage indicates more repetitive text; lower suggests richer vocabulary.

### Word Embedding Metrics

**Isotropy**
> *Definition:* Measures how uniformly distributed vectors are in the embedding space. Computed as the ratio of minimum to maximum singular values.
>
> *Intuition:* High isotropy (near 1.0) means vectors spread evenly in all directions; low isotropy means vectors cluster in certain directions, reducing expressiveness.
>
> *What to seek:* Higher isotropy generally indicates better-quality embeddings. Values > 0.1 are reasonable; > 0.3 is good. Lower-dimensional embeddings tend to have higher isotropy.

**Average Norm**
> *Definition:* Mean magnitude (L2 norm) of word vectors in the embedding space.
>
> *Intuition:* Indicates the typical "length" of vectors. Consistent norms suggest stable training; high variance may indicate some words are undertrained.
>
> *What to seek:* Relatively consistent norms across models. The absolute value matters less than consistency (low std deviation).

**Cosine Similarity**
> *Definition:* Measures angular similarity between vectors, ranging from -1 (opposite) to 1 (identical direction).
>
> *Intuition:* Words with similar meanings should have high cosine similarity. This is the standard metric for semantic relatedness in embeddings.
>
> *What to seek:* Semantically related words should score > 0.5; unrelated words should be near 0. Synonyms often score > 0.7.

**t-SNE Visualization**
> *Definition:* t-Distributed Stochastic Neighbor Embedding - a dimensionality reduction technique that preserves local structure for visualization.
>
> *Intuition:* Clusters in t-SNE plots indicate groups of semantically related words. Spread indicates vocabulary diversity; tight clusters suggest semantic coherence.
>
> *What to seek:* Meaningful clusters (e.g., numbers together, verbs together). Avoid over-interpreting distances - t-SNE preserves local, not global, structure.

### General Interpretation Guidelines

1. **Compare within model families:** Metrics are most meaningful when comparing models of the same type (e.g., 8k vs 64k tokenizer).
2. **Consider trade-offs:** Better performance on one metric often comes at the cost of another (e.g., compression vs. OOV rate).
3. **Context matters:** Optimal values depend on downstream tasks. Text generation may prioritize different metrics than classification.
4. **Corpus influence:** All metrics are influenced by corpus characteristics. Wikipedia text differs from social media or literature.
5. **Language-specific patterns:** Morphologically rich languages (like Arabic) may show different optimal ranges than analytic languages.


### Visualizations Index

| Visualization | Description |
|---------------|-------------|
| Tokenizer Compression | Compression ratios by vocabulary size |
| Tokenizer Fertility | Average token length by vocabulary |
| Tokenizer OOV | Unknown token rates |
| Tokenizer Total Tokens | Total tokens by vocabulary |
| N-gram Perplexity | Perplexity by n-gram size |
| N-gram Entropy | Entropy by n-gram size |
| N-gram Coverage | Top pattern coverage |
| N-gram Unique | Unique n-gram counts |
| Markov Entropy | Entropy by context size |
| Markov Branching | Branching factor by context |
| Markov Contexts | Unique context counts |
| Zipf's Law | Frequency-rank distribution with fit |
| Vocab Frequency | Word frequency distribution |
| Top 20 Words | Most frequent words |
| Vocab Coverage | Cumulative coverage curve |
| Embedding Isotropy | Vector space uniformity |
| Embedding Norms | Vector magnitude distribution |
| Embedding Similarity | Word similarity heatmap |
| Nearest Neighbors | Similar words for key terms |
| t-SNE Words | 2D word embedding visualization |
| t-SNE Sentences | 2D sentence embedding visualization |
| Position Encoding | Encoding method comparison |
| Model Sizes | Storage requirements |
| Performance Dashboard | Comprehensive performance overview |

---
## About This Project

### Data Source

Models trained on [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly) - a monthly snapshot of Wikipedia articles across 300+ languages.

### Project

A project by **[Wikilangs](https://wikilangs.org)** - Open-source NLP models for every Wikipedia language.

### Maintainer

[Omar Kamali](https://omarkamali.com) - [Omneity Labs](https://omneitylabs.com)

### Citation

If you use these models in your research, please cite:

```bibtex
@misc{wikilangs2025,
  author = {Kamali, Omar},
  title = {Wikilangs: Open NLP Models for Wikipedia Languages},
  year = {2025},
  doi = {10.5281/zenodo.18073153},
  publisher = {Zenodo},
  url = {https://huggingface.co/wikilangs}
  institution = {Omneity Labs}
}
```

### License

MIT License - Free for academic and commercial use.

### Links

- 🌐 Website: [wikilangs.org](https://wikilangs.org)
- 🤗 Models: [huggingface.co/wikilangs](https://huggingface.co/wikilangs)
- 📊 Data: [wikipedia-monthly](https://huggingface.co/datasets/omarkamali/wikipedia-monthly)
- 👤 Author: [Omar Kamali](https://huggingface.co/omarkamali)
- 🤝 Sponsor: [Featherless AI](https://featherless.ai)
---
*Generated by Wikilangs Models Pipeline*

*Report Date: 2026-01-10 22:33:22*