README.md

---
language: bjn
language_name: Banjar
language_family: austronesian_malay
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-austronesian_malay
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: 4.830
  - name: best_isotropy
    type: isotropy
    value: 0.8715
  - name: vocabulary_size
    type: vocab
    value: 0
generated: 2026-01-03
---

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

This repository contains NLP models trained and evaluated by Wikilangs, specifically on **Banjar** 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** | 3.761x | 3.76 | 0.3950% | 367,048 |
| **16k** | 4.164x | 4.17 | 0.4374% | 331,539 |
| **32k** | 4.537x | 4.54 | 0.4766% | 304,229 |
| **64k** | 4.830x 🏆 | 4.83 | 0.5073% | 285,820 |

### Tokenization Examples

Below are sample sentences tokenized with each vocabulary size:

**Sample 1:** `Wedoro adalah sabuah kampung di Kacamatan Glagah, Kabupatin Lamongan, Prupinsi J...`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁w ed oro ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁glagah , ... (+9 more)` | 19 |
| 16k | `▁wed oro ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁glagah , ▁kabupatin ... (+8 more)` | 18 |
| 32k | `▁wedoro ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁glagah , ▁kabupatin ▁lamongan ... (+7 more)` | 17 |
| 64k | `▁wedoro ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁glagah , ▁kabupatin ▁lamongan ... (+7 more)` | 17 |

**Sample 2:** `Laburan Baru' adalah sabuah kampung di Kacamatan Paser Belengkong, Kabupatin Pas...`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁lab uran ▁baru ' ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁paser ... (+12 more)` | 22 |
| 16k | `▁lab uran ▁baru ' ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁paser ... (+11 more)` | 21 |
| 32k | `▁laburan ▁baru ' ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁paser ▁belengkong ... (+10 more)` | 20 |
| 64k | `▁laburan ▁baru ' ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁paser ▁belengkong ... (+10 more)` | 20 |

**Sample 3:** `Nibung adalah sabuah kampung di Kacamatan Selimbau, Kabupatin Kapuas Hulu, Prupi...`

| Vocab | Tokens | Count |
|-------|--------|-------|
| 8k | `▁n ib ung ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁sel imb ... (+12 more)` | 22 |
| 16k | `▁n ibung ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁selimbau , ▁kabupatin ... (+9 more)` | 19 |
| 32k | `▁nibung ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁selimbau , ▁kabupatin ▁kapuas ... (+8 more)` | 18 |
| 64k | `▁nibung ▁adalah ▁sabuah ▁kampung ▁di ▁kacamatan ▁selimbau , ▁kabupatin ▁kapuas ... (+8 more)` | 18 |


### Key Findings

- **Best Compression:** 64k achieves 4.830x compression
- **Lowest UNK Rate:** 8k with 0.3950% 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 | 6,505 | 12.67 | 21,751 | 23.6% | 44.5% |
| **2-gram** | Subword | 185 🏆 | 7.53 | 2,788 | 78.2% | 99.5% |
| **3-gram** | Word | 3,849 | 11.91 | 17,881 | 32.5% | 51.6% |
| **3-gram** | Subword | 1,428 | 10.48 | 20,293 | 34.4% | 80.3% |
| **4-gram** | Word | 5,302 | 12.37 | 24,831 | 28.9% | 48.0% |
| **4-gram** | Subword | 7,612 | 12.89 | 99,642 | 17.5% | 50.0% |
| **5-gram** | Word | 4,712 | 12.20 | 16,656 | 25.7% | 48.4% |
| **5-gram** | Subword | 25,009 | 14.61 | 245,459 | 12.3% | 34.0% |

### Top 5 N-grams by Size

**2-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `kampung di` | 5,961 |
| 2 | `prupinsi kalimantan` | 5,903 |
| 3 | `di kacamatan` | 5,625 |
| 4 | `adalah sabuah` | 4,211 |
| 5 | `sabuah kampung` | 3,806 |

**3-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `kampung di kacamatan` | 5,212 |
| 2 | `sabuah kampung di` | 3,803 |
| 3 | `adalah sabuah kampung` | 3,803 |
| 4 | `kalimantan selatan indunisia` | 2,201 |
| 5 | `prupinsi kalimantan selatan` | 2,188 |

**4-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `sabuah kampung di kacamatan` | 3,802 |
| 2 | `adalah sabuah kampung di` | 3,801 |
| 3 | `prupinsi kalimantan selatan indunisia` | 2,154 |
| 4 | `prupinsi kalimantan barat indunisia` | 1,806 |
| 5 | `yaitu sabuting kampung di` | 1,356 |

**5-grams (Word):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `adalah sabuah kampung di kacamatan` | 3,801 |
| 2 | `yaitu sabuting kampung di kacamatan` | 1,253 |
| 3 | `indunisia géografi watas wilayah watas` | 1,113 |
| 4 | `géografi watas wilayah watas wilayah` | 1,099 |
| 5 | `watas wilayah watas wilayah kacamatan` | 739 |

**2-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `a n` | 365,243 |
| 2 | `n _` | 194,875 |
| 3 | `n g` | 152,971 |
| 4 | `a _` | 138,836 |
| 5 | `k a` | 132,349 |

**3-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `a n _` | 156,222 |
| 2 | `a n g` | 84,871 |
| 3 | `_ k a` | 76,502 |
| 4 | `n g _` | 75,610 |
| 5 | `_ m a` | 57,961 |

**4-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `a n g _` | 48,934 |
| 2 | `t a n _` | 34,621 |
| 3 | `n a n g` | 29,979 |
| 4 | `a t a n` | 29,470 |
| 5 | `_ n a n` | 28,658 |

**5-grams (Subword):**

| Rank | N-gram | Count |
|------|--------|-------|
| 1 | `_ n a n g` | 28,407 |
| 2 | `n a n g _` | 27,864 |
| 3 | `a t a n _` | 22,485 |
| 4 | `m a t a n` | 17,997 |
| 5 | `_ w a n _` | 17,178 |


### Key Findings

- **Best Perplexity:** 2-gram (subword) with 185
- **Entropy Trend:** Decreases with larger n-grams (more predictable)
- **Coverage:** Top-1000 patterns cover ~34% 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 | 0.8469 | 1.799 | 5.67 | 99,056 | 15.3% |
| **1** | Subword | 0.7172 | 1.644 | 4.59 | 2,416 | 28.3% |
| **2** | Word | 0.2337 | 1.176 | 1.48 | 559,810 | 76.6% |
| **2** | Subword | 0.6823 | 1.605 | 4.16 | 11,092 | 31.8% |
| **3** | Word | 0.0561 | 1.040 | 1.08 | 824,984 | 94.4% |
| **3** | Subword | 0.7802 | 1.717 | 3.90 | 46,118 | 22.0% |
| **4** | Word | 0.0150 🏆 | 1.010 | 1.02 | 890,043 | 98.5% |
| **4** | Subword | 0.6544 | 1.574 | 2.81 | 179,736 | 34.6% |

### Generated Text Samples (Word-based)

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

**Context Size 1:**

1. `di kacamatan konang kabupatin sanggau prupinsi kalimantan tengah mesir india indunisia watas wilayah...`
2. `nang baisi banyak banar dalam bahasa utama liga 3 m 1 sampai pamulaan wan takananya barupa`
3. `wan manangani kajahatan gasan hintalu diploid buhannya kawa jua gasan pahitungan hisab nitu angin tu...`

**Context Size 2:**

1. `kampung di kacamatan teluk sampit pambagian administratip kacamatan tualan hulu pambagian administra...`
2. `prupinsi kalimantan timur indunisia makanan nangkaya tempe matan kacang kacangan imbah disangrai bad...`
3. `di kacamatan menyuke kabupatin landak prupinsi kalimantan barat indunisia géografi watas wilayah kac...`

**Context Size 3:**

1. `kampung di kacamatan tambakrejo kabupatin bojonegoro prupinsi jawa timur jujuhutan`
2. `adalah sabuah kampung di kacamatan semitau kabupatin kapuas hulu prupinsi kalimantan barat indunisia...`
3. `sabuah kampung di kacamatan long iram kabupatin kutai barat prupinsi kalimantan timur indunisia géog...`

**Context Size 4:**

1. `sabuah kampung di kacamatan bengalon kabupatin kutai timur prupinsi kalimantan timur indunisia indun...`
2. `adalah sabuah kampung di kacamatan ketungau tengah kabupatin sintang prupinsi kalimantan barat indun...`
3. `yaitu sabuting kampung di kacamatan karang intan kabupatin banjar prupinsi kalimantan selatan induni...`


### Generated Text Samples (Subword-based)

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

**Context Size 1:**

1. `awanaangik_ta,_t`
2. `_g_viabara_pa_li`
3. `ng_ksawarbunteru`

**Context Size 2:**

1. `anyan_adangga,_br`
2. `n_kalambang_pem_a`
3. `ng_dew,_dibantu,_`

**Context Size 3:**

1. `an_jejani_andan_ka`
2. `ang_sambara,_pres,`
3. `_kacamatas_palima_`

**Context Size 4:**

1. `ang_maman_banjadi_h`
2. `tan_bakcanganis_rik`
3. `nang_kampung_dalah_`


### Key Findings

- **Best Predictability:** Context-4 (word) with 98.5% predictability
- **Branching Factor:** Decreases with context size (more deterministic)
- **Memory Trade-off:** Larger contexts require more storage (179,736 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 | 41,351 |
| Total Tokens | 992,449 |
| Mean Frequency | 24.00 |
| Median Frequency | 4 |
| Frequency Std Dev | 278.70 |

### Most Common Words

| Rank | Word | Frequency |
|------|------|-----------|
| 1 | di | 27,655 |
| 2 | nang | 27,387 |
| 3 | wan | 17,250 |
| 4 | adalah | 10,715 |
| 5 | lawan | 9,581 |
| 6 | indunisia | 9,420 |
| 7 | kacamatan | 9,139 |
| 8 | kalimantan | 8,368 |
| 9 | kampung | 7,824 |
| 10 | matan | 7,698 |

### Least Common Words (from vocabulary)

| Rank | Word | Frequency |
|------|------|-----------|
| 1 | beregszásziová | 2 |
| 2 | košice | 2 |
| 3 | satian | 2 |
| 4 | extreme | 2 |
| 5 | frisna | 2 |
| 6 | ropang | 2 |
| 7 | caknan | 2 |
| 8 | muktamar | 2 |
| 9 | sandon | 2 |
| 10 | sékuéns | 2 |

### Zipf's Law Analysis

| Metric | Value |
|--------|-------|
| Zipf Coefficient | 1.0491 |
| R² (Goodness of Fit) | 0.995109 |
| Adherence Quality | **excellent** |

### Coverage Analysis

| Top N Words | Coverage |
|-------------|----------|
| Top 100 | 35.5% |
| Top 1,000 | 62.4% |
| Top 5,000 | 81.6% |
| Top 10,000 | 88.8% |

### Key Findings

- **Zipf Compliance:** R²=0.9951 indicates excellent adherence to Zipf's law
- **High Frequency Dominance:** Top 100 words cover 35.5% of corpus
- **Long Tail:** 31,351 words needed for remaining 11.2% 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 | 0.8715 | 0.3303 | N/A | N/A |
| **mono_64d** | 64 | 0.8409 | 0.2593 | N/A | N/A |
| **mono_128d** | 128 | 0.5527 | 0.2130 | N/A | N/A |
| **aligned_32d** | 32 | 0.8715 🏆 | 0.3312 | 0.0420 | 0.2520 |
| **aligned_64d** | 64 | 0.8409 | 0.2582 | 0.0680 | 0.3160 |
| **aligned_128d** | 128 | 0.5527 | 0.2256 | 0.1380 | 0.4260 |

### Key Findings

- **Best Isotropy:** aligned_32d with 0.8715 (more uniform distribution)
- **Semantic Density:** Average pairwise similarity of 0.2696. Lower values indicate better semantic separation.
- **Alignment Quality:** Aligned models achieve up to 13.8% 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 | **0.423** | 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 |
|--------|----------|
| `-ma` | manentang, maut, marked |
| `-pa` | parachute, pattern, pamain |
| `-ba` | bantam, babakan, barambai |
| `-di` | dibawakan, dihimpun, dibatasi |
| `-ka` | karoseri, kampanye, kahala |
| `-ta` | tatikap, tahitung, tato |
| `-man` | manentang, manuruti, manggalungsur |
| `-pe` | penyelenggara, pengadilan, pertapaan |

#### Productive Suffixes
| Suffix | Examples |
|--------|----------|
| `-n` | pattern, babakan, tikinan |
| `-an` | babakan, tikinan, kanaan |
| `-a` | kurbannya, kahala, dhaka |
| `-ng` | manentang, gondang, rahang |
| `-kan` | babakan, dibawakan, menguntungkan |
| `-ya` | kurbannya, karibnya, makanannya |
| `-nya` | kurbannya, karibnya, makanannya |
| `-akan` | babakan, dibawakan, maruntuhakan |

### 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` | 1.62x | 225 contexts | sanga, manga, nanga |
| `unga` | 2.11x | 57 contexts | bunga, rungan, bungas |
| `ngan` | 1.95x | 58 contexts | pangan, rungan, bongan |
| `anja` | 1.76x | 82 contexts | sanja, ganja, anjat |
| `ntan` | 1.89x | 49 contexts | antan, intan, antang |
| `mant` | 1.94x | 39 contexts | manta, manti, mantel |
| `ting` | 1.63x | 79 contexts | keting, tingah, eating |
| `ndun` | 2.15x | 24 contexts | rundun, indung, mendung |
| `dala` | 1.77x | 38 contexts | dalam, dalas, adalah |
| `atin` | 1.82x | 26 contexts | atina, batin, latin |
| `pung` | 1.91x | 21 contexts | apung, pungsi, capung |
| `adal` | 1.91x | 16 contexts | badal, kadal, adalah |

### 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 |
|--------|--------|-----------|----------|
| `-pa` | `-n` | 207 words | palayanan, paampihan |
| `-pa` | `-an` | 195 words | palayanan, paampihan |
| `-di` | `-n` | 149 words | diasingakan, dimanangakan |
| `-ma` | `-n` | 144 words | manyurangan, mampartahanakan |
| `-ka` | `-n` | 144 words | kamantirian, kajiwaan |
| `-di` | `-an` | 140 words | diasingakan, dimanangakan |
| `-ma` | `-an` | 136 words | manyurangan, mampartahanakan |
| `-di` | `-kan` | 133 words | diasingakan, dimanangakan |
| `-ka` | `-an` | 133 words | kamantirian, kajiwaan |
| `-ma` | `-kan` | 126 words | mampartahanakan, maungkaiakan |

### 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 |
|------|-----------------|------------|------|
| kaputingannya | **`ka-puti-ng-an-nya`** | 9.0 | `puti` |
| dimanpaatakan | **`di-man-pa-atak-an`** | 9.0 | `atak` |
| manjadiakannya | **`man-jadi-akan-nya`** | 7.5 | `jadi` |
| mamandiakan | **`ma-man-di-akan`** | 7.5 | `akan` |
| disayangakan | **`di-sa-yang-akan`** | 7.5 | `yang` |
| peradangan | **`pe-rada-ng-an`** | 7.5 | `rada` |
| dimakamakan | **`di-ma-ka-makan`** | 7.5 | `makan` |
| kakacangan | **`ka-ka-cang-an`** | 7.5 | `cang` |
| disalanggarakan | **`di-sa-langgar-akan`** | 7.5 | `langgar` |
| takapinggirakan | **`ta-ka-pinggir-akan`** | 7.5 | `pinggir` |
| dihasilakannya | **`di-hasil-akan-nya`** | 7.5 | `hasil` |
| pahitungan | **`pa-hitu-ng-an`** | 7.5 | `hitu` |
| papadahannya | **`pa-pa-dahan-nya`** | 7.5 | `dahan` |
| sabalumannya | **`sa-ba-luman-nya`** | 7.5 | `luman` |
| kahiringan | **`ka-hiri-ng-an`** | 7.5 | `hiri` |

### 6.6 Linguistic Interpretation

> **Automated Insight:**
The language Banjar 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 | **64k BPE** | Best compression (4.83x) |
| N-gram | **2-gram** | Lowest perplexity (185) |
| Markov | **Context-4** | Highest predictability (98.5%) |
| 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 19:11:59*