README.md

---
language: chy
language_name: Cheyenne
language_family: american_algonquian
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-american_algonquian
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: 3.494
  - name: best_isotropy
    type: isotropy
    value: 0.0023
  - name: vocabulary_size
    type: vocab
    value: 0
generated: 2026-01-03
---

# Cheyenne - Wikilangs Models
## Comprehensive Research Report & Full Ablation Study

This repository contains NLP models trained and evaluated by Wikilangs, specifically on **Cheyenne** 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** | 3.494x 🏆 | 3.52 | 0.1022% | 18,598 |

### Tokenization Examples

Below are sample sentences tokenized with each vocabulary size:

**Sample 1:** `Vóo'kooma, vóo'ooma (Melanerpes erythrocephalus) ve'kêseho-éve. Tôhohko`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁vóo ' kooma , ▁vóo ' ooma ▁( melanerpes ▁erythrocephalus ... (+8 more)` | 18 |

**Sample 2:** `Hestaahtsémeno (Ribes floridum), heso'xêhestaahtsémeno, na'éstse máhtáme.`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁hestaahtsémeno ▁( ribes ▁floridum ), ▁heso ' xêhestaahtsémeno , ▁na ... (+4 more)` | 14 |

**Sample 3:** `Vó'aehesanestôtse (vé'ho'énêstsestôtse: buckskin suit; "antelope-dress") Pl: vó'...`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁vó ' aehesanestôtse ▁( vé ' ho ' énêstsestôtse : ... (+20 more)` | 30 |


### Key Findings

- **Best Compression:** 8k achieves 3.494x compression
- **Lowest UNK Rate:** 8k with 0.1022% 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 | 98 🏆 | 6.62 | 148 | 88.0% | 100.0% |
| **2-gram** | Subword | 325 | 8.34 | 853 | 59.8% | 100.0% |
| **3-gram** | Word | 150 | 7.23 | 229 | 74.0% | 100.0% |
| **3-gram** | Subword | 1,635 | 10.67 | 3,634 | 27.6% | 73.9% |
| **4-gram** | Word | 301 | 8.23 | 420 | 52.7% | 100.0% |
| **4-gram** | Subword | 3,873 | 11.92 | 8,064 | 18.7% | 53.5% |
| **5-gram** | Word | 213 | 7.74 | 290 | 59.9% | 100.0% |
| **5-gram** | Subword | 4,512 | 12.14 | 8,516 | 17.1% | 49.3% |

### Top 5 N-grams by Size

**2-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `na éstse` | 140 |
| 2 | `vé ho` | 119 |
| 3 | `ho énêstsestôtse` | 72 |
| 4 | `republic of` | 67 |
| 5 | `éstse manâhéno` | 55 |

**3-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `vé ho énêstsestôtse` | 72 |
| 2 | `na éstse manâhéno` | 55 |
| 3 | `ho honáéšé e` | 44 |
| 4 | `ho e éve` | 33 |
| 5 | `éstse ho e` | 32 |

**4-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `na éstse ho e` | 32 |
| 2 | `éstse ho e éve` | 32 |
| 3 | `ma kaetaévôxe êstoo o` | 25 |
| 4 | `toháano éve ho etse` | 23 |
| 5 | `manâhéno ho honáéšé e` | 22 |

**5-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `na éstse ho e éve` | 32 |
| 2 | `ho honáéšé e united states` | 22 |
| 3 | `éstse manâhéno ho honáéšé e` | 22 |
| 4 | `na éstse manâhéno ho honáéšé` | 22 |
| 5 | `manâhéno ho honáéšé e united` | 21 |

**2-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `e _` | 1,450 |
| 2 | `s e` | 1,334 |
| 3 | `s t` | 1,269 |
| 4 | `t s` | 1,249 |
| 5 | `h e` | 974 |

**3-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `t s e` | 956 |
| 2 | `s e _` | 548 |
| 3 | `e s t` | 461 |
| 4 | `s t s` | 436 |
| 5 | `h o '` | 420 |

**4-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `t s e _` | 427 |
| 2 | `s t s e` | 413 |
| 3 | `ô t s e` | 276 |
| 4 | `t ô t s` | 204 |
| 5 | `e s t ô` | 194 |

**5-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `s t s e _` | 216 |
| 2 | `t ô t s e` | 203 |
| 3 | `s t ô t s` | 190 |
| 4 | `e s t ô t` | 190 |
| 5 | `ê s t s e` | 170 |


### Key Findings

- **Best Perplexity:** 2-gram (word) with 98
- **Entropy Trend:** Decreases with larger n-grams (more predictable)
- **Coverage:** Top-1000 patterns cover ~49% 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 | 0.4049 | 1.324 | 1.97 | 3,214 | 59.5% |
| **1** | Subword | 1.3402 | 2.532 | 9.42 | 172 | 0.0% |
| **2** | Word | 0.1099 | 1.079 | 1.20 | 6,126 | 89.0% |
| **2** | Subword | 1.2169 | 2.324 | 5.05 | 1,620 | 0.0% |
| **3** | Word | 0.0453 | 1.032 | 1.08 | 7,065 | 95.5% |
| **3** | Subword | 0.6471 | 1.566 | 2.32 | 8,158 | 35.3% |
| **4** | Word | 0.0256 🏆 | 1.018 | 1.04 | 7,317 | 97.4% |
| **4** | Subword | 0.2799 | 1.214 | 1.44 | 18,852 | 72.0% |

### Generated Text Samples (Word-based)

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

**Context Size 1:**

1. `e éve ho honáéšé e cfa ma kaetaévôxe êstoo o toháano éve hóxovê hooma naa kánome`
2. `ho éstova éhe nėstaane néstse vóonotse 30 hestáotse naa unie van zuid afrika hotómá e great`
3. `o gdp ppp 72 7 afrikaans vé ho etse 56 785 6 coloured 9 indian tséh`

**Context Size 2:**

1. `na éstse manâhéno ho honáéšé e vehicle license kȧhkoetohko prefix 29 hotómá e mo hetaneho e hánêsóvó...`
2. `vé ho énestse 71 740 6 144 562 903 somali federal republic of the congo congo kinshasa`
3. `ho énêstsestôtse wyolacheyenne english dictionarychief dull knife college hoig stan the peace chiefs...`

**Context Size 3:**

1. `vé ho énêstsestôtse airplane this is`
2. `na éstse manâhéno china republic of china republic of china republic of china republic of china repu...`
3. `ho honáéšé e native news project`

**Context Size 4:**

1. `éstse ho e éve vietnam dong hoi airport`
2. `na éstse ho e éve united states states of america`
3. `ma kaetaévôxe êstoo o toháano éve ho etse 322 460 1 600 democratic republic of the congo of the`


### Generated Text Samples (Subword-based)

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

**Context Size 1:**

1. `etokfive_piente'`
2. `_t:_manésé'e'e,_`
3. `aliotse'étinoo's`

**Context Size 2:**

1. `e_100px_minestȯts`
2. `se_cre_manéó'ho'ô`
3. `stanjunt.thumb_la`

**Context Size 3:**

1. `tse_(lephonáéšé'e,`
2. `se_odom_capid_city`
3. `estôtsestôtsestôts`

**Context Size 4:**

1. `tse_évȯhkėha'etaneh`
2. `stsestȯtse_kóhkonôh`
3. `ôtsenáesëö'o_môxeov`


### Key Findings

- **Best Predictability:** Context-4 (word) with 97.4% predictability
- **Branching Factor:** Decreases with context size (more deterministic)
- **Memory Trade-off:** Larger contexts require more storage (18,852 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 | 1,174 |
| Total Tokens | 7,828 |
| Mean Frequency | 6.67 |
| Median Frequency | 3 |
| Frequency Std Dev | 21.01 |

### Most Common Words

| Rank | Word | Frequency |
|------|------|-----------|
| 1 | e | 407 |
| 2 | ho | 351 |
| 3 | o | 229 |
| 4 | vé | 159 |
| 5 | na | 144 |
| 6 | éstse | 140 |
| 7 | éve | 133 |
| 8 | of | 117 |
| 9 | naa | 104 |
| 10 | he | 103 |

### Least Common Words (from vocabulary)

| Rank | Word | Frequency |
|------|------|-----------|
| 1 | pack | 2 |
| 2 | evenóse | 2 |
| 3 | mountain | 2 |
| 4 | cal | 2 |
| 5 | poly | 2 |
| 6 | mustangs | 2 |
| 7 | sevonévo | 2 |
| 8 | ėstovátamevéotse | 2 |
| 9 | ėstova | 2 |
| 10 | nėstse | 2 |

### Zipf's Law Analysis

| Metric | Value |
|--------|-------|
| Zipf Coefficient | 0.8142 |
| R² (Goodness of Fit) | 0.973597 |
| Adherence Quality | **excellent** |

### Coverage Analysis

| Top N Words | Coverage |
|-------------|----------|
| Top 100 | 55.3% |
| Top 1,000 | 95.6% |
| Top 5,000 | 0.0% |
| Top 10,000 | 0.0% |

### Key Findings

- **Zipf Compliance:** R²=0.9736 indicates excellent adherence to Zipf's law
- **High Frequency Dominance:** Top 100 words cover 55.3% of corpus
- **Long Tail:** -8,826 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 | 0.0023 🏆 | 0.8896 | N/A | N/A |
| **mono_64d** | 64 | 0.0007 | 0.9590 | N/A | N/A |
| **mono_128d** | 128 | 0.0002 | 0.9907 | N/A | N/A |
| **aligned_32d** | 32 | 0.0023 | 0.8896 | 0.0513 | 0.2179 |
| **aligned_64d** | 64 | 0.0007 | 0.9590 | 0.0385 | 0.1795 |
| **aligned_128d** | 128 | 0.0002 | 0.9907 | 0.0128 | 0.1667 |

### Key Findings

- **Best Isotropy:** mono_32d with 0.0023 (more uniform distribution)
- **Semantic Density:** Average pairwise similarity of 0.9464. Lower values indicate better semantic separation.
- **Alignment Quality:** Aligned models achieve up to 5.1% 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 | **1.027** | 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 |
|--------|----------|
| `-ho` | hotóao, hohtóvá, hoéstónéó |

#### Productive Suffixes
| Suffix | Examples |
|--------|----------|
| `-e` | ôhkêhenove, háahpe, manâhestôtse |
| `-se` | manâhestôtse, tsétsêhéstâhese, xaénéhetse |
| `-tse` | manâhestôtse, xaénéhetse, ôhnéménêstse |
| `-ôtse` | manâhestôtse, mâhoestôtse, ôtse |
| `-ne` | lione, mâhoestôtsene, nemâhmoteone |
| `-ve` | ôhkêhenove, ôhkemôxeonêstove, kêsaéve |
| `-ia` | alnifolia, austria, nitsvia |

### 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.

*No significant bound stems detected.*


### 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 |
|--------|--------|-----------|----------|
| `-ho` | `-e` | 5 words | house, hovahne |
| `-ho` | `-ne` | 2 words | hovahne, hovane |
| `-ho` | `-se` | 1 words | house, hotse |
| `-ho` | `-tse` | 1 words | hotse, hohpâhtsenámenôtse |
| `-ho` | `-ôtse` | 1 words | hohpâhtsenámenôtse |

### 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 |
|------|-----------------|------------|------|
| mâhoestôtsene | **`mâhoest-ôtse-ne`** | 3.0 | `mâhoest` |
| sevoneóneve | **`sevoneó-ne-ve`** | 3.0 | `sevoneó` |
| éestsėstóseoneve | **`éestsėstóseo-ne-ve`** | 3.0 | `éestsėstóseo` |
| enóseoneve | **`enóseo-ne-ve`** | 3.0 | `enóseo` |
| náhkȯhehetanetse | **`náhkȯheheta-ne-tse`** | 3.0 | `náhkȯheheta` |
| ôhkêhenove | **`ôhkêheno-ve`** | 1.5 | `ôhkêheno` |
| manâhestôtse | **`manâhest-ôtse`** | 1.5 | `manâhest` |
| alnifolia | **`alnifol-ia`** | 1.5 | `alnifol` |
| ôhkemôxeonêstove | **`ôhkemôxeonêsto-ve`** | 1.5 | `ôhkemôxeonêsto` |
| hoéstónéó | **`ho-éstónéó`** | 1.5 | `éstónéó` |
| nemâhmoteone | **`nemâhmoteo-ne`** | 1.5 | `nemâhmoteo` |
| tsétsêhéstâhese | **`tsétsêhéstâhe-se`** | 1.5 | `tsétsêhéstâhe` |
| australia | **`austral-ia`** | 1.5 | `austral` |
| shepherdia | **`shepherd-ia`** | 1.5 | `shepherd` |
| xaénéhetse | **`xaénéhe-tse`** | 1.5 | `xaénéhe` |

### 6.6 Linguistic Interpretation

> **Automated Insight:**
The language Cheyenne 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 | **8k BPE** | Best compression (3.49x) |
| N-gram | **2-gram** | Lowest perplexity (98) |
| Markov | **Context-4** | Highest predictability (97.4%) |
| 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: 2026-01-03 20:28:03*