Upload all models and assets for arc (latest)

.gitattributes

@@ -39,3 +39,4 @@ visualizations/position_encoding_comparison.png filter=lfs diff=lfs merge=lfs -t
 visualizations/tsne_sentences.png filter=lfs diff=lfs merge=lfs -text
 visualizations/tsne_words.png filter=lfs diff=lfs merge=lfs -text
 visualizations/zipf_law.png filter=lfs diff=lfs merge=lfs -text
+visualizations/embedding_tsne_multilingual.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,6 +1,6 @@
 ---
 language: arc
-language_name: ARC
+language_name: Aramaic
 language_family: semitic_aramaic
 tags:
   - wikilangs
@@ -10,11 +10,21 @@ tags:
   - n-gram
   - markov
   - wikipedia
+  - feature-extraction
+  - sentence-similarity
+  - tokenization
+  - n-grams
+  - markov-chain
+  - text-mining
+  - fasttext
+  - babelvec
+  - vocabulous
+  - vocabulary
   - monolingual
   - family-semitic_aramaic
 license: mit
 library_name: wikilangs
-pipeline_tag: feature-extraction
+pipeline_tag: text-generation
 datasets:
   - omarkamali/wikipedia-monthly
 dataset_info:
@@ -26,17 +36,17 @@ metrics:
     value: 4.583
   - name: best_isotropy
     type: isotropy
-    value: 0.2739
+    value: 0.3326
   - name: vocabulary_size
     type: vocab
     value: 0
 generated: 2026-01-03
 ---
 
-# ARC - Wikilangs Models
+# Aramaic - Wikilangs Models
 ## Comprehensive Research Report & Full Ablation Study
 
-This repository contains NLP models trained and evaluated by Wikilangs, specifically on **ARC** Wikipedia data.
+This repository contains NLP models trained and evaluated by Wikilangs, specifically on **Aramaic** Wikipedia data.
 We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and word embeddings.
 
 ## 📋 Repository Contents
@@ -60,7 +70,7 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
 - [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)
+- [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)
@@ -80,43 +90,43 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
 
 | Vocab Size | Compression | Avg Token Len | UNK Rate | Total Tokens |
 |------------|-------------|---------------|----------|--------------|
-| **8k** | 3.552x | 3.57 | 0.1271% | 63,747 |
-| **16k** | 3.988x | 4.01 | 0.1427% | 56,780 |
-| **32k** | 4.583x 🏆 | 4.60 | 0.1640% | 49,402 |
+| **8k** | 3.553x | 3.57 | 0.1262% | 63,406 |
+| **16k** | 3.990x | 4.01 | 0.1417% | 56,456 |
+| **32k** | 4.583x 🏆 | 4.60 | 0.1628% | 49,148 |
 
 ### Tokenization Examples
 
 Below are sample sentences tokenized with each vocabulary size:
 
-**Sample 1:** `ܡܬܠܬܐ ܗܘ ܐܣܟܡܐ ܡܚܪܝܐ (ܓܐܘܡܛܪܝܐ) ܕܐܝܬ ܠܗ ܬܠܬܐ ܐܠܥ̈ܐ ܘܬܠܬܐ ܙܘܝܬ̈ܐ܀`
+**Sample 1:** `ܐܫܬܐ (ܟܢܘܫܝܐ: ܐܫܝ̈ܬܐ) ܗܝ ܡܢܬܐ ܓܠܝܠܬܐ ܕܨܪܘܝܘܬܐ ܕܬܦܠ ܒܒܣܬܪܐ ܕܐܓܢܐ ܕܓܘܫܡ̈ܐ ܕܐܢܫ̈ܐ ܘ...`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 8k | `▁ܡܬܠܬܐ ▁ܗܘ ▁ܐܣܟܡܐ ▁ܡܚܪܝܐ ▁( ܓܐ ܘܡ ܛܪܝܐ ) ▁ܕܐܝܬ ... (+8 more)` | 18 |
-| 16k | `▁ܡܬܠܬܐ ▁ܗܘ ▁ܐܣܟܡܐ ▁ܡܚܪܝܐ ▁( ܓܐܘܡܛܪܝܐ ) ▁ܕܐܝܬ ▁ܠܗ ▁ܬܠܬܐ ... (+4 more)` | 14 |
-| 32k | `▁ܡܬܠܬܐ ▁ܗܘ ▁ܐܣܟܡܐ ▁ܡܚܪܝܐ ▁( ܓܐܘܡܛܪܝܐ ) ▁ܕܐܝܬ ▁ܠܗ ▁ܬܠܬܐ ... (+3 more)` | 13 |
+| 8k | `▁ܐܫܬܐ ▁( ܟܢܘܫܝܐ : ▁ܐ ܫܝ ̈ܬܐ ) ▁ܗܝ ▁ܡܢܬܐ ... (+20 more)` | 30 |
+| 16k | `▁ܐܫܬܐ ▁( ܟܢܘܫܝܐ : ▁ܐ ܫܝ ̈ܬܐ ) ▁ܗܝ ▁ܡܢܬܐ ... (+18 more)` | 28 |
+| 32k | `▁ܐܫܬܐ ▁( ܟܢܘܫܝܐ : ▁ܐܫܝ̈ܬܐ ) ▁ܗܝ ▁ܡܢܬܐ ▁ܓܠܝܠܬܐ ▁ܕܨܪܘܝܘܬܐ ... (+12 more)` | 22 |
 
-**Sample 2:** `ܟܐܢܣܐܣ ܐܘ ܟܐܢܙܐܣ (Kansas) ܐܝܬܝܗ ܐܘܚܕܢܐ ܓܘ ܡܢܬܐ ܡܥܪܒܝܬܐ ܡܨܥܝܬܐ ܕܐ̈ܘܚܕܢܐ ܡ̈ܚܝܕܐ ܕܐ...`
+**Sample 2:** `ܐܫܬܐ ܗܝ ܓܕܫܐ ܐܣܝܝܐ. ܐܫܬܐ ܗܝ ܚܡܝܡܘܬ ܓܘܫܡܐ ܠܥܠ ܡܢ ܫܘܝܐ ܟܝܢܝܐ ܕܚܡܝܡܘܬܐ ܓܘܝܬܐ ܕܓܘܫܡܐ...`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 8k | `▁ܟ ܐܢ ܣܐܣ ▁ܐܘ ▁ܟ ܐܢ ܙܐ ܣ ▁( k ... (+14 more)` | 24 |
-| 16k | `▁ܟܐܢܣܐܣ ▁ܐܘ ▁ܟܐܢܙܐܣ ▁( kansas ) ▁ܐܝܬܝܗ ▁ܐܘܚܕܢܐ ▁ܓܘ ▁ܡܢܬܐ ... (+7 more)` | 17 |
-| 32k | `▁ܟܐܢܣܐܣ ▁ܐܘ ▁ܟܐܢܙܐܣ ▁( kansas ) ▁ܐܝܬܝܗ ▁ܐܘܚܕܢܐ ▁ܓܘ ▁ܡܢܬܐ ... (+7 more)` | 17 |
+| 8k | `▁ܐܫܬܐ ▁ܗܝ ▁ܓܕܫܐ ▁ܐܣ ܝܝܐ . ▁ܐܫܬܐ ▁ܗܝ ▁ܚܡ ܝܡ ... (+10 more)` | 20 |
+| 16k | `▁ܐܫܬܐ ▁ܗܝ ▁ܓܕܫܐ ▁ܐܣܝܝܐ . ▁ܐܫܬܐ ▁ܗܝ ▁ܚܡܝܡܘܬ ▁ܓܘܫܡܐ ▁ܠܥܠ ... (+6 more)` | 16 |
+| 32k | `▁ܐܫܬܐ ▁ܗܝ ▁ܓܕܫܐ ▁ܐܣܝܝܐ . ▁ܐܫܬܐ ▁ܗܝ ▁ܚܡܝܡܘܬ ▁ܓܘܫܡܐ ▁ܠܥܠ ... (+6 more)` | 16 |
 
-**Sample 3:** `ܢܝܘ ܗܐܡܦܫܪ (New Hampshire) ܗܝ ܐܬܪܐ ܓܘ ܡܢܬܐ ܓܪܒܝܝܬܐ ܡܕܢܚܝܬܐ ܕܐܬ݂ܪ̈ܘܬ݂ܐ ܡ̈ܚܝܕܐ ܕܐܡ...`
+**Sample 3:** `ܡܪܝ ܢܪܣܝ ܕܒܙ (ܐܬܝܠܕ 17 ܐܝܪ - ܡܝܬ 14 ܫܒܛ ܗܘܐ ܡܝܛܪܦܘܠܝܛܐ ܕܠܒܢܢ ܘܣܘܪܝܐ ܘܟܠܗ̇ ܐܘܪܘܦܐ...`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 8k | `▁ܢܝܘ ▁ܗܐܡ ܦܫ ܪ ▁( new ▁h am p shi ... (+14 more)` | 24 |
-| 16k | `▁ܢܝܘ ▁ܗܐܡܦܫܪ ▁( new ▁hamp shire ) ▁ܗܝ ▁ܐܬܪܐ ▁ܓܘ ... (+9 more)` | 19 |
-| 32k | `▁ܢܝܘ ▁ܗܐܡܦܫܪ ▁( new ▁hampshire ) ▁ܗܝ ▁ܐܬܪܐ ▁ܓܘ ▁ܡܢܬܐ ... (+8 more)` | 18 |
+| 8k | `▁ܡܪܝ ▁ܢܪܣܝ ▁ܕܒܙ ▁( ܐܬܝܠܕ ▁ 1 7 ▁ܐܝܪ ▁- ... (+17 more)` | 27 |
+| 16k | `▁ܡܪܝ ▁ܢܪܣܝ ▁ܕܒܙ ▁( ܐܬܝܠܕ ▁ 1 7 ▁ܐܝܪ ▁- ... (+17 more)` | 27 |
+| 32k | `▁ܡܪܝ ▁ܢܪܣܝ ▁ܕܒܙ ▁( ܐܬܝܠܕ ▁ 1 7 ▁ܐܝܪ ▁- ... (+16 more)` | 26 |
 
 
 ### Key Findings
 
 - **Best Compression:** 32k achieves 4.583x compression
-- **Lowest UNK Rate:** 8k with 0.1271% unknown tokens
+- **Lowest UNK Rate:** 8k with 0.1262% unknown tokens
 - **Trade-off:** Larger vocabularies improve compression but increase model size
 - **Recommendation:** 32k vocabulary provides optimal balance for production use
 
@@ -133,12 +143,14 @@ Below are sample sentences tokenized with each vocabulary size:
 
 | N-gram | Variant | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
 |--------|---------|------------|---------|----------------|------------------|-------------------|
-| **2-gram** | Word | 477 | 8.90 | 718 | 45.8% | 100.0% |
-| **2-gram** | Subword | 365 🏆 | 8.51 | 2,347 | 59.8% | 96.1% |
-| **3-gram** | Word | 437 | 8.77 | 752 | 52.0% | 100.0% |
-| **3-gram** | Subword | 2,390 | 11.22 | 10,625 | 28.2% | 66.9% |
-| **4-gram** | Word | 742 | 9.53 | 1,438 | 43.7% | 84.6% |
-| **4-gram** | Subword | 8,576 | 13.07 | 28,979 | 14.2% | 43.1% |
+| **2-gram** | Word | 477 | 8.90 | 719 | 45.8% | 100.0% |
+| **2-gram** | Subword | 363 🏆 | 8.50 | 2,330 | 59.9% | 96.1% |
+| **3-gram** | Word | 438 | 8.77 | 754 | 52.0% | 100.0% |
+| **3-gram** | Subword | 2,379 | 11.22 | 10,583 | 28.3% | 67.0% |
+| **4-gram** | Word | 759 | 9.57 | 1,466 | 43.3% | 83.8% |
+| **4-gram** | Subword | 8,542 | 13.06 | 28,872 | 14.3% | 43.2% |
+| **5-gram** | Word | 508 | 8.99 | 1,055 | 50.5% | 97.5% |
+| **5-gram** | Subword | 16,168 | 13.98 | 38,919 | 8.4% | 30.9% |
 
 ### Top 5 N-grams by Size
 
@@ -148,66 +160,86 @@ Below are sample sentences tokenized with each vocabulary size:
 |------|--------|-------|
 | 1 | `ܐܦ ܚܙܝ` | 193 |
 | 2 | `ܚܕ ܡܢ` | 141 |
-| 3 | `ܗܝ ܐܬܪܐ` | 123 |
-| 4 | `ܐܝܬ ܠܗ` | 103 |
-| 5 | `ܬܚܘܡܐ ܥܡ` | 88 |
+| 3 | `ܗܝ ܐܬܪܐ` | 124 |
+| 4 | `ܐܝܬ ܠܗ` | 102 |
+| 5 | `ܬܚܘܡܐ ܥܡ` | 89 |
 
 **3-grams (Word):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
 | 1 | `ܗܘ ܚܕ ܡܢ` | 72 |
-| 2 | `ܢܕܢܐ ܡܠܝܝܐ ܢܟܓܐܝܚܢܛܟ‍` | 52 |
-| 3 | `ܡܒܕ ܫܐܢܡܝܢ ܪܡܝܚܢܐܢ` | 52 |
-| 4 | `ܒܟܡ ܣܢܝܓܚܝܢܪܢ ܟܢܫܙܢ` | 52 |
-| 5 | `ܣܢܝܓܚܝܢܪܢ ܟܢܫܙܢ ܢܝܛܠܐ` | 52 |
+| 2 | `ܟܠܢܚܝܓܐ ܓܐ ܝܢܦܠ` | 52 |
+| 3 | `ܝܠܟܐܒ ܝܓܚܝܐ ܟܠܢܚܝܓܐ` | 52 |
+| 4 | `ܝܢܦܠ ܡܒܤܢ ܐܤܡ` | 52 |
+| 5 | `ܓܐ ܝܢܦܠ ܡܒܤܢ` | 52 |
 
 **4-grams (Word):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `ܐܤܡ ܟܛܠ ܚܢܝܬܝܐ ܡܕܛܚܝܢܐ` | 52 |
-| 2 | `ܢܝܛܠܐ ܝܟܝܟܕ ܝܡܓܚܝܢܐ ܐܓܐ` | 52 |
-| 3 | `ܝܟܝܟܕ ܝܡܓܚܝܢܐ ܐܓܐ ܟܡܠܐ` | 52 |
-| 4 | `ܝܡܓܚܝܢܐ ܐܓܐ ܟܡܠܐ ܣܐܙܬܝܐܢ` | 52 |
-| 5 | `ܡܓܝܡܡ ܡܟܒܡ ܠܣܐܟ ܒܟܡ` | 52 |
+| 1 | `ܝܡܓܚܝܢܐ ܐܓܐ ܟܡܠܐ ܣܐܙܬܝܐܢ` | 52 |
+| 2 | `ܝܟܝܟܕ ܝܡܓܚܝܢܐ ܐܓܐ ܟܡܠܐ` | 52 |
+| 3 | `ܢܝܛܠܐ ܝܟܝܟܕ ܝܡܓܚܝܢܐ ܐܓܐ` | 52 |
+| 4 | `ܟܢܫܙܢ ܢܝܛܠܐ ܝܟܝܟܕ ܝܡܓܚܝܢܐ` | 52 |
+| 5 | `ܣܢܝܓܚܝܢܪܢ ܟܢܫܙܢ ܢܝܛܠܐ ܝܟܝܟܕ` | 52 |
+
+**5-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ܝܠܟܐܒ ܝܓܚܝܐ ܟܠܢܚܝܓܐ ܓܐ ܝܢܦܠ` | 52 |
+| 2 | `ܟܠܢܚܝܓܐ ܓܐ ܝܢܦܠ ܡܒܤܢ ܐܤܡ` | 52 |
+| 3 | `ܓܐ ܝܢܦܠ ܡܒܤܢ ܐܤܡ ܟܛܠ` | 52 |
+| 4 | `ܡܒܤܢ ܐܤܡ ܟܛܠ ܚܢܝܬܝܐ ܡܕܛܚܝܢܐ` | 52 |
+| 5 | `ܝܢܦܠ ܡܒܤܢ ܐܤܡ ܟܛܠ ܚܢܝܬܝܐ` | 52 |
 
 **2-grams (Subword):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `ܐ _` | 24,633 |
-| 2 | `_ ܕ` | 7,621 |
-| 3 | `ܬ ܐ` | 7,176 |
-| 4 | `_ ܐ` | 6,899 |
-| 5 | `ܝ ܐ` | 5,702 |
+| 1 | `ܐ _` | 24,552 |
+| 2 | `_ ܕ` | 7,580 |
+| 3 | `ܬ ܐ` | 7,166 |
+| 4 | `_ ܐ` | 6,890 |
+| 5 | `ܝ ܐ` | 5,689 |
 
 **3-grams (Subword):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `ܐ _ ܕ` | 6,138 |
-| 2 | `ܬ ܐ _` | 5,890 |
-| 3 | `ܝ ܐ _` | 4,242 |
+| 1 | `ܐ _ ܕ` | 6,099 |
+| 2 | `ܬ ܐ _` | 5,875 |
+| 3 | `ܝ ܐ _` | 4,233 |
 | 4 | `ܐ _ ܐ` | 2,477 |
-| 5 | `ܢ ܐ _` | 2,397 |
+| 5 | `ܐ _ ܘ` | 2,392 |
 
 **4-grams (Subword):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `ܬ ܐ _ ܕ` | 2,008 |
-| 2 | `ܝ ܬ ܐ _` | 1,523 |
-| 3 | `ܐ ܝ ܬ _` | 1,367 |
-| 4 | `ܘ ܬ ܐ _` | 1,297 |
-| 5 | `_ ܡ ܢ _` | 1,210 |
+| 1 | `ܬ ܐ _ ܕ` | 1,993 |
+| 2 | `ܝ ܬ ܐ _` | 1,513 |
+| 3 | `ܐ ܝ ܬ _` | 1,372 |
+| 4 | `ܘ ܬ ܐ _` | 1,304 |
+| 5 | `_ ܡ ܢ _` | 1,203 |
+
+**5-grams (Subword):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `ܐ _ ܐ ܘ _` | 603 |
+| 2 | `ܘ ܬ ܐ _ ܕ` | 543 |
+| 3 | `ܐ _ ܡ ܢ _` | 533 |
+| 4 | `_ ܐ ܝ ܬ _` | 503 |
+| 5 | `ܝ ܘ ܬ ܐ _` | 487 |
 
 
 ### Key Findings
 
-- **Best Perplexity:** 2-gram (subword) with 365
+- **Best Perplexity:** 2-gram (subword) with 363
 - **Entropy Trend:** Decreases with larger n-grams (more predictable)
-- **Coverage:** Top-1000 patterns cover ~43% of corpus
+- **Coverage:** Top-1000 patterns cover ~31% of corpus
 - **Recommendation:** 4-gram or 5-gram for best predictive performance
 
 ---
@@ -223,14 +255,14 @@ Below are sample sentences tokenized with each vocabulary size:
 
 | Context | Variant | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
 |---------|---------|-------------|------------|------------------|-----------------|----------------|
-| **1** | Word | 0.5449 | 1.459 | 2.60 | 18,018 | 45.5% |
-| **1** | Subword | 0.9655 | 1.953 | 6.06 | 1,232 | 3.5% |
-| **2** | Word | 0.1027 | 1.074 | 1.16 | 45,749 | 89.7% |
-| **2** | Subword | 0.7977 | 1.738 | 3.85 | 7,459 | 20.2% |
-| **3** | Word | 0.0295 | 1.021 | 1.04 | 51,822 | 97.1% |
-| **3** | Subword | 0.5934 | 1.509 | 2.45 | 28,618 | 40.7% |
-| **4** | Word | 0.0106 🏆 | 1.007 | 1.01 | 52,472 | 98.9% |
-| **4** | Subword | 0.3583 | 1.282 | 1.71 | 69,915 | 64.2% |
+| **1** | Word | 0.5444 | 1.458 | 2.59 | 17,962 | 45.6% |
+| **1** | Subword | 0.9651 | 1.952 | 6.05 | 1,231 | 3.5% |
+| **2** | Word | 0.1025 | 1.074 | 1.16 | 45,549 | 89.7% |
+| **2** | Subword | 0.7954 | 1.736 | 3.84 | 7,433 | 20.5% |
+| **3** | Word | 0.0296 | 1.021 | 1.04 | 51,591 | 97.0% |
+| **3** | Subword | 0.5935 | 1.509 | 2.45 | 28,465 | 40.7% |
+| **4** | Word | 0.0106 🏆 | 1.007 | 1.01 | 52,240 | 98.9% |
+| **4** | Subword | 0.3585 | 1.282 | 1.71 | 69,540 | 64.2% |
 
 ### Generated Text Samples (Word-based)
 
@@ -238,27 +270,27 @@ Below are text samples generated from each word-based Markov chain model:
 
 **Context Size 1:**
 
-1. `ܡܢ ܓܪܒܝܐ ܕܒܝܬ ܗܘܠܢܕ̈ܝܐ ܗܘܠܢܕܐܝܬ hengelo ܗܝ ܐܘܚܕܢܐ ܒܓܪܒܝܐ ܕܥܝܪܐܩ ܝܘܡܢܐ ܬܡܢ ܣܝܡܠܗ̇ ܥܕܬ̈ܐ ܡܕܢܚܝܬ̈ܐ ܕܐܬ݂...`
-2. `ܐܘ ܙܢܓܒܝܠܐ ܗܘ ܡܣܘܪܩܐ ܡܐܢܐ ܕܐܪܕܟܠܘܬܐ`
-3. `ܗܘ ܓܘܣܐ ܒܥܠܬܐ ܐܘ ܝܣܪܝܠ ܐܘ ܬܫܪܝܢ ܒ ܩܛܠܥܡܐ ܣܘܪܝܝܐ ܒ ت ܬ ܦ 80 754`
+1. `ܡܢ ܕܝܪܐ ܕܡܪܝ ܓܒܪܐܝܠ ܗܘ ܫܡܐ ܕܒܗܘ ܢܬܝܕܥܘܢ ܗܘܘ ܡܠܘܢ ܕܝܢ ܪ̈ܥܘܬܐ ܕܝܠܗܘܢ ܒܪܓܘܠܐ ܕܓܗܢܐ ܗܢܐ`
+2. `ܐܘ ܡܣܐܢܐ ܐܘ ܚܡܗܐ ܗܘ ܣܦܪܐ ܕܗܘܫܥ ܘܕܐܫܥܝܐ ܘܗܘܝܘ ܦܪܝܣܐ ܩܘܢܣܛܢܛܝܢ`
+3. `ܗܘ ܢܒܝܐ ܙܥܘܪܐ ܒܠܒܘܫܐ ܕܒܗ ܫܩܠܘ ܐܢܛܝܘܟܝܐ ܘܫܚܠܦ ܗܘܐ ܦܛܪܝܪܟܐ ܩܢܘܢܝܐ ܣܕܪܐ ܘܟܠ ܕܡܘ 2 ܒܢܝܣܢ`
 
 **Context Size 2:**
 
-1. `ܐܦ ܚܙܝ ܓܒܪܐ`
-2. `ܚܕ ܡܢ ܐܪܒܥܐ ܟܬܒ̈ܐ ܩܕ̈ܡܝܐ ܕܕܝܬܝܩܝ ܚܕܬܐ ܦܘܠܘܣ ܫܠܝܚܐ ܟܬܒ ܗܕܐ ܐܓܪܬܐ ܠܩܘܠܣܝ̈ܐ ܕܗܢܘܢ ܐܢܫ̈ܐ ܕܡܕܝܢܬܐ ܕܐܦܣܘܣ`
-3. `ܗܝ ܐܬܪܐ ܒܐܘܪܘܦܐ ܩܘܛܢܝܘܬܐ ܕܐܝܪܠܢܕ ܗܝ ܒܓܘ ܓܙܪܬܐ ܕܐܝܪܠܢܕ ܠܐܝܪܠܢܕ ܓܪܒܝܝܬܐ ܐܝܬ ܬܚܘܡܐ ܥܡ ܪܘܡܢܝܐ ܘܥܡ ܛܘܪܩܝܐ`
+1. `ܐܦ ܚܙܝ ܥܡܐ ܒܝܬܝܘܬܐ de verwandtschaftsbeziehung onkel und tante`
+2. `ܚܕ ܡܢ ܬܪܥܣܪ ܢܒܝ̈ܐ ܙܥܘܪ̈ܐ ܕܬܢܟ ܘܕܕܝܬܝܩܝ ܥܬܝܩܬܐ ܥܬܝܩܬܐ`
+3. `ܗܝ ܐܬܪܐ ܒܐܣܝܐ ܨܝܢ ܐܝܬ ܠܗ̇ ܪܡܙܐ ܕܡܘܠܕܐ ܬܪܝܢܐ ܒܝܕ ܡܝ̈ܐ ܘܪܘܚܐ ܕܩܘܕܫܐ ܝܫܘܥ ܐܡܪ ܠܗ ܘܠܐܚܗ`
 
 **Context Size 3:**
 
-1. `ܗܘ ܚܕ ܡܢ ܓܘܢ̈ܐ ܪ̈ܝܫܝܐ ܕܗܢܘܢ ܣܘܡܩܐ ܘܫܥܘܬܐ ܘܙܪܩܐ ܢܘܗܪܐ ܣܘܡܩܐ ܐܝܬ ܠܗ ܐܘܪܟܐ ܓܠܠܝܐ ܢܐܢܘܡܝܛܪ`
-2. `ܡܓܝܡܡ ܡܟܒܡ ܠܣܐܟ ܒܟܡ ܣܢܝܓܚܝܢܪܢ ܟܢܫܙܢ ܢܝܛܠܐ ܝܟܝܟܕ ܝܡܓܚܝܢܐ ܐܓܐ ܟܡܠܐ ܣܐܙܬܝܐܢ ܝܠܟܐܒ ܝܓܚܝܐ ܟܠܢܚܝܓܐ ܓܐ ܝܢܦܠ...`
-3. `ܡܕܛܚܝܢܐ ܡܒܕ ܫܐܢܡܝܢ ܪܡܝܚܢܐܢ ܢܕܢܐ ܡܠܝܝܐ ܢܟܓܐܝܚܢܛܟ‍ ܟ‍ܝܣܢܐ ܡܓܝܡܡ ܡܟܒܡ ܠܣܐܟ ܒܟܡ ܣܢܝܓܚܝܢܪܢ ܟܢܫܙܢ ܢܝܛܠܐ ܝܟ...`
+1. `ܗܘ ܚܕ ܡܢ ܠܫܢ̈ܐ ܨܝܢܝ̈ܐ ܕܢܬܡܠܠܘܢ ܒܬܝܡܢ ܡܕܢܚ ܕܨܝܢ ܐܝܬܘܗܝ ܠܫܢܐ ܐܡܗܝܐ ܕܝܬܝܪ ܡܢ 90 ܡܠܝܘܢܐ ܐܢܫ̈ܝܢ ܪܝܫܐܝܬ`
+2. `ܢܝܛܠܐ ܝܟܝܟܕ ܝܡܓܚܝܢܐ ܐܓܐ ܟܡܠܐ ܣܐܙܬܝܐܢ ܝܠܟܐܒ ܝܓܚܝܐ ܟܠܢܚܝܓܐ ܓܐ ܝܢܦܠ ܡܒܤܢ ܐܤܡ ܟܛܠ ܚܢܝܬܝܐ ܡܕܛܚܝܢܐ ܡܒܕ ܫܐܢ...`
+3. `ܟܛܠ ܚܢܝܬܝܐ ܡܕܛܚܝܢܐ ܡܒܕ ܫܐܢܡܝܢ ܪܡܝܚܢܐܢ ܢܕܢܐ ܡܠܝܝܐ ܢܟܓܐܝܚܢܛܟ‍ ܟ‍ܝܣܢܐ ܡܓܝܡܡ ܡܟܒܡ ܠܣܐܟ ܒܟܡ ܣܢܝܓܚܝܢܪܢ ܟܢܫ...`
 
 **Context Size 4:**
 
-1. `ܡܒܕ ܫܐܢܡܝܢ ܪܡܝܚܢܐܢ ܢܕܢܐ ܡܠܝܝܐ ܢܟܓܐܝܚܢܛܟ‍ ܟ‍ܝܣܢܐ ܡܓܝܡܡ ܡܟܒܡ ܠܣܐܟ ܒܟܡ ܣܢܝܓܚܝܢܪܢ ܟܢܫܙܢ ܢܝܛܠܐ ܝܟܝܟܕ ܝܡܓܚ...`
-2. `ܡܓܝܡܡ ܡܟܒܡ ܠܣܐܟ ܒܟܡ ܣܢܝ��ܚܝܢܪܢ ܟܢܫܙܢ ܢܝܛܠܐ ܝܟܝܟܕ ܝܡܓܚܝܢܐ ܐܓܐ ܟܡܠܐ ܣܐܙܬܝܐܢ ܝܠܟܐܒ ܝܓܚܝܐ ܟܠܢܚܝܓܐ ܓܐ ܝܢܦܠ...`
-3. `ܝܠܟܐܒ ܝܓܚܝܐ ܟܠܢܚܝܓܐ ܓܐ ܝܢܦܠ ܡܒܤܢ ܐܤܡ ܟܛܠ ܚܢܝܬܝܐ ܡܕܛܚܝܢܐ ܡܒܕ ܫܐܢܡܝܢ ܪܡܝܚܢܐܢ ܢܕܢܐ ܡܠܝܝܐ ܢܟܓܐܝܚܢܛܟ‍ ܟ‍ܝ...`
+1. `ܝܓܚܝܐ ܟܠܢܚܝܓܐ ܓܐ ܝܢܦܠ ܡܒܤܢ ܐܤܡ ܟܛܠ ܚܢܝܬܝܐ ܡܕܛܚܝܢܐ ܡܒܕ ܫܐܢܡܝܢ ܪܡܝܚܢܐܢ ܢܕܢܐ ܡܠܝܝܐ ܢܟܓܐܝܚܢܛܟ‍ ܟ‍ܝܣܢܐ ܡܓ...`
+2. `ܚܢܝܬܝܐ ܡܕܛܚܝܢܐ ܡܒܕ ܫܐܢܡܝܢ ܪܡܝܚܢܐܢ ܢܕܢܐ ܡܠܝܝܐ ܢܟܓܐܝܚܢܛܟ‍ ܟ‍ܝܣܢܐ ܡܓܝܡܡ ܡܟܒܡ ܠܣܐܟ ܒܟܡ ܣܢܝܓܚܝܢܪܢ ܟܢܫܙܢ ܢ...`
+3. `ܟܛܠ ܚܢܝܬܝܐ ܡܕܛܚܝܢܐ ܡܒܕ ܫܐܢܡܝܢ ܪܡܝܚܢܐܢ ܢܕܢܐ ܡܠܝܝܐ ܢܟܓܐܝܚܢܛܟ‍ ܟ‍ܝܣܢܐ ܡܓܝܡܡ ܡܟܒܡ ܠܣܐܟ ܒܟܡ ܣܢܝܓܚܝܢܪܢ ܟܢܫ...`
 
 
 ### Generated Text Samples (Subword-based)
@@ -267,34 +299,34 @@ Below are text samples generated from each subword-based Markov chain model:
 
 **Context Size 1:**
 
-1. `_ܨܝܬܐ_أبيد)_ܙܘܢܓ`
-2. `ܐ._ܥܡܫܝܗܝܢܝܬ̈ܐ_ܡܕ`
-3. `ܝܟܫܬܐ_ܐ܀_ܐ_ܫܘܪܒܡ`
+1. `_ܥܪ_ܙܥܠܐܡܘܦܐܘܗ_ܡ`
+2. `ܐܣܬܐ_ܒܕ_ܐ_ܡܐ_ܕܡܦ`
+3. `ܝܓܝܛܠܚܕܢܪܝܦܝ̈ܘܪܝܐ`
 
 **Context Size 2:**
 
-1. `ܐ_ܕܗܘܡܝܠܢܕܐ_ܕܐܥܬܐ`
-2. `_ܕܥܣܪܘܝܕܝܢܘܢ_ܐܘ_ܦ`
-3. `ܬܐ_ܕܩܪܝܬܐ_ܥܠ_ܐܟܪܝ`
+1. `ܐ_ܫܡܐ_ܕܢܐܡܪܝܬ_800`
+2. `_ܕܠܐܬܘܕܝܫܐ_ܘܒܝܬ_ا`
+3. `ܬܐ_ܕܛܚܝܬ_ܡܚܝܬ_ܡܥܘ`
 
 **Context Size 3:**
 
-1. `ܐ_ܕܒܓܕܐ_ܕܛܘܪ_ܗܘܘ_ܬ`
-2. `ܬܐ_ܕܛܲܟ݂ܣܵܐ_ܡܠܟܐ_ܫܠܝܚ̈`
-3. `ܝܐ_ܘܡܬܝ_ܡܠܝܝܐ_מלזי`
+1. `ܐ_ܕܐܬܥܢܘܬܐ_ܠܫܢܐ_(r`
+2. `ܬܐ_ܥܪܒܐ_ܩܘܛܢܝܘܬܐ_ܛ`
+3. `ܝܐ_ܒܪ_ܒܪ_ܐܘ_ܐܘܢܛܐܪ`
 
 **Context Size 4:**
 
-1. `ܬܐ_ܕܗܘܦܪܟܝܐ_ܕܪܗܘܡܝܬ`
-2. `ܝܬܐ_ܙܘ_ܫܘܬܐ_ܕܬܘܕܝܬܐ`
-3. `ܐܝܬ_ܡܨܪ̈ܝܐ܆_ܚܣܢ_ܒܡܕܝ`
+1. `ܬܐ_ܕܢܒܥ_ܡܫܝܚܐ._ܥܠܠܢ̈`
+2. `ܝܬܐ_ܕܪ̈ܗܘܡܝܐ܀_ܢܗܪܝܢ܁`
+3. `ܐܝܬ_ܚܡܫܐ_ܨܒܝܢܗ._ܟܬܒ`
 
 
 ### Key Findings
 
 - **Best Predictability:** Context-4 (word) with 98.9% predictability
 - **Branching Factor:** Decreases with context size (more deterministic)
-- **Memory Trade-off:** Larger contexts require more storage (69,915 contexts)
+- **Memory Trade-off:** Larger contexts require more storage (69,540 contexts)
 - **Recommendation:** Context-3 or Context-4 for text generation
 
 ---
@@ -310,9 +342,9 @@ Below are text samples generated from each subword-based Markov chain model:
 
 | Metric | Value |
 |--------|-------|
-| Vocabulary Size | 6,113 |
-| Total Tokens | 50,830 |
-| Mean Frequency | 8.32 |
+| Vocabulary Size | 6,099 |
+| Total Tokens | 50,661 |
+| Mean Frequency | 8.31 |
 | Median Frequency | 3 |
 | Frequency Std Dev | 32.05 |
 
@@ -320,16 +352,16 @@ Below are text samples generated from each subword-based Markov chain model:
 
 | Rank | Word | Frequency |
 |------|------|-----------|
-| 1 | ܡܢ | 1,283 |
-| 2 | ܐܘ | 975 |
-| 3 | ܗܘ | 861 |
+| 1 | ܡܢ | 1,276 |
+| 2 | ܐܘ | 979 |
+| 3 | ܗܘ | 860 |
 | 4 | ܗܝ | 816 |
-| 5 | ܐܝܬ | 512 |
+| 5 | ܐܝܬ | 513 |
 | 6 | ܗܘܐ | 394 |
-| 7 | ܥܠ | 327 |
-| 8 | ܘܥܡ | 326 |
-| 9 | ܐܦ | 275 |
-| 10 | ܠܫܢܐ | 266 |
+| 7 | ܘܥܡ | 330 |
+| 8 | ܥܠ | 324 |
+| 9 | ܐܦ | 277 |
+| 10 | ܠܫܢܐ | 263 |
 
 ### Least Common Words (from vocabulary)
 
@@ -350,24 +382,24 @@ Below are text samples generated from each subword-based Markov chain model:
 
 | Metric | Value |
 |--------|-------|
-| Zipf Coefficient | 0.8947 |
-| R² (Goodness of Fit) | 0.982774 |
+| Zipf Coefficient | 0.8942 |
+| R² (Goodness of Fit) | 0.982775 |
 | Adherence Quality | **excellent** |
 
 ### Coverage Analysis
 
 | Top N Words | Coverage |
 |-------------|----------|
-| Top 100 | 31.7% |
+| Top 100 | 31.8% |
 | Top 1,000 | 68.0% |
-| Top 5,000 | 95.6% |
+| Top 5,000 | 95.7% |
 | Top 10,000 | 0.0% |
 
 ### Key Findings
 
 - **Zipf Compliance:** R²=0.9828 indicates excellent adherence to Zipf's law
-- **High Frequency Dominance:** Top 100 words cover 31.7% of corpus
-- **Long Tail:** -3,887 words needed for remaining 100.0% coverage
+- **High Frequency Dominance:** Top 100 words cover 31.8% of corpus
+- **Long Tail:** -3,901 words needed for remaining 100.0% coverage
 
 ---
 ## 5. Word Embeddings Evaluation
@@ -383,37 +415,40 @@ Below are text samples generated from each subword-based Markov chain model:
 
 ### 5.1 Cross-Lingual Alignment
 
-> *Note: Multilingual alignment visualization not available for this language.*
+![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 | 0.2739 🏆 | 0.4979 | N/A | N/A |
-| **mono_64d** | 64 | 0.0566 | 0.5064 | N/A | N/A |
-| **mono_128d** | 128 | 0.0089 | 0.4882 | N/A | N/A |
+| **mono_32d** | 32 | 0.3326 | 0.4816 | N/A | N/A |
+| **mono_64d** | 64 | 0.0524 | 0.4997 | N/A | N/A |
+| **mono_128d** | 128 | 0.0094 | 0.4954 | N/A | N/A |
+| **aligned_32d** | 32 | 0.3326 🏆 | 0.4744 | 0.2099 | 0.5556 |
+| **aligned_64d** | 64 | 0.0524 | 0.4826 | 0.1975 | 0.6543 |
+| **aligned_128d** | 128 | 0.0094 | 0.5048 | 0.2099 | 0.7037 |
 
 ### Key Findings
 
-- **Best Isotropy:** mono_32d with 0.2739 (more uniform distribution)
-- **Semantic Density:** Average pairwise similarity of 0.4975. Lower values indicate better semantic separation.
-- **Alignment Quality:** No aligned models evaluated in this run.
+- **Best Isotropy:** aligned_32d with 0.3326 (more uniform distribution)
+- **Semantic Density:** Average pairwise similarity of 0.4897. Lower values indicate better semantic separation.
+- **Alignment Quality:** Aligned models achieve up to 21.0% R@1 in cross-lingual retrieval.
 - **Recommendation:** 128d aligned for best cross-lingual performance
 
 ---
 ## 6.  Morphological Analysis (Experimental)
 
-> ⚠️ **Warning:** This language shows low morphological productivity. The statistical signals used for this analysis may be noisy or less reliable than for morphologically rich languages.
-
 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 | **0.000** | Low morphological productivity | ⚠️ Likely unreliable |
-| Idiomaticity Gap | **-1.000** | Low formulaic content | - |
+| Productivity Index | **5.000** | High morphological productivity | Reliable analysis |
+| Idiomaticity Gap | **2.160** | High formulaic/idiomatic content | - |
 
 ### 6.2 Affix Inventory (Productive Units)
 
@@ -426,13 +461,13 @@ These are the most productive prefixes and suffixes identified by sampling the v
 #### Productive Suffixes
 | Suffix | Examples |
 |--------|----------|
-| `-ܐ` | ܘܫܛܚܐ, ܠܘܚ̈ܐ, ܐܢܫܘܬܐ |
-| `-ܬܐ` | ܐܢܫܘܬܐ, ܚܘܪܬܐ, ܐܘܪܬܕܘܟܣܝܬܐ |
-| `-ܝܐ` | ܘܥܪ̈ܒܝܐ, ܒܐܠܒܢܝܐ, ܥܒܝܐ |
-| `-̈ܐ` | ܠܘܚ̈ܐ, ܡܚܝܕ̈ܐ, ܥܝܢ̈ܐ |
-| `-ܝܬܐ` | ܐܘܪܬܕܘܟܣܝܬܐ, ܣܘܪܝܝܬܐ, ܒܩܕܡܝܬܐ |
-| `-ܘܬܐ` | ܐܢܫܘܬܐ, ܘܕܡܠܟܘܬܐ, ܛܝܒܘܬܐ |
-| `-ܢܐ` | ܕܫܝܢܐ, ܡܢܝܢܐ, ܥܝܢܐ |
+| `-ܐ` | ܬܫܐܢܐܟܐܠܐ, ܟܘܕܢܝܐ, ܕܚܘܼܡܵܐ |
+| `-ܬܐ` | ܐܒܪ̈ܗܡܝܬܐ, ܕܡܥܡܕܝܬܐ, ܫܬܐ |
+| `-ܝܐ` | ܟܘܕܢܝܐ, ܡܠܝܝܐ, ܓܐܘܓܪܦܝܐ |
+| `-ܝܬܐ` | ܐܒܪ̈ܗܡܝܬܐ, ܕܡܥܡܕܝܬܐ, ܟܢܘܫܝܬܐ |
+| `-̈ܐ` | ܢܒܝ̈ܐ, ܕܓܘܫܡ̈ܐ, ܘܚܝ̈ܐ |
+| `-ܘܬܐ` | ܘܐܬܪ̈ܘܬܐ, ܛܝܒܘܬܐ, ܦܛܪܝܪܟܘܬܐ |
+| `-ܢܐ` | ܐܡܝܢܐ, ܘܡܩܝܡܢܐ, ܕܓܝܗܢܐ |
 
 ### 6.3 Bound Stems (Lexical Roots)
 
@@ -440,18 +475,18 @@ Bound stems are high-frequency subword units that are semantically cohesive but
 
 | Stem | Cohesion | Substitutability | Examples |
 |------|----------|------------------|----------|
-| `ܢܝܬܐ` | 1.58x | 23 contexts | ܦܢܝܬܐ, ܡܢܝܬܐ, ܡܪܢܝܬܐ |
-| `ܪܝܬܐ` | 1.59x | 18 contexts | ܫܪܝܬܐ, ܩܪܝܬܐ, ܒܪܝܬܐ |
-| `ܫܝܚܝ` | 1.61x | 16 contexts | ܡܫܝܚܝܐ, ܡܫܝܚܝܬܐ, ܕܡܫܝܚܝܐ |
-| `ܪܒܝܐ` | 1.59x | 16 contexts | ܨܪܒܝܐ, ܥܪܒܝܐ, ܐܪܒܝܐ |
-| `ܘܢܝܐ` | 1.57x | 16 contexts | ܟܘܢܝܐ, ܩܘܢܝܐ, ܓܘܢܝܐ |
-| `ܘܪܝܐ` | 1.37x | 23 contexts | ܛܘܪܝܐ, ܣܘܪܝܐ, ܟܘܪܝܐ |
-| `ܡܫܝܚ` | 1.59x | 14 contexts | ܡܫܝܚܐ, ܡܫܝܚܝܐ, ܕܡܫܝܚܐ |
-| `ܡܕܝܢ` | 1.58x | 13 contexts | ܡܕܝܢܬ, ܡܕܝܢܬܐ, ܠܡܕܝܢܬ |
-| `ܢܐܝܬ` | 1.53x | 14 contexts | ܝܘܢܐܝܬ, ܨܝܢܐܝܬ, ܟܠܢܐܝܬ |
-| `ܣܘܪܝ` | 1.38x | 18 contexts | ܣܘܪܝܐ, ܣܘܪܝܬ, ܘܣܘܪܝܐ |
-| `ܝܢܬܐ` | 1.65x | 9 contexts | ܩܝܢܬܐ, ܡܕܝܢܬܐ, ܣܦܝܢܬܐ |
-| `ܕܝܢܬ` | 1.62x | 9 contexts | ܡܕܝܢܬ, ܡܕܝܢܬܐ, ܠܡܕܝܢܬ |
+| `ܢܝܬܐ` | 1.64x | 23 contexts | ܦܢܝܬܐ, ܡܢܝܬܐ, ܡܕܢܝܬܐ |
+| `ܪܝܬܐ` | 1.68x | 18 contexts | ܒܪܝܬܐ, ܫܪܝܬܐ, ܩܪܝܬܐ |
+| `ܘܪܝܐ` | 1.51x | 23 contexts | ܣܘܪܝܐ, ܛܘܪܝܐ, ܟܘܪܝܐ |
+| `ܫܝܚܝ` | 1.69x | 15 contexts | ܡܫܝܚܝܐ, ܘܡܫܝܚܝܐ, ܡܫܝܚܝ̈ܐ |
+| `ܪܒܝܐ` | 1.63x | 16 contexts | ܨܪܒܝܐ, ܓܪܒܝܐ, ܐܪܒܝܐ |
+| `ܘܢܝܐ` | 1.63x | 15 contexts | ܝܘܢܝܐ, ܓܘܢܝܐ, ܟܘܢܝܐ |
+| `ܡܫܝܚ` | 1.70x | 13 contexts | ܡܫܝܚܐ, ܡܫܝܚܝܐ, ܕܡܫܝܚܐ |
+| `ܣܘܪܝ` | 1.49x | 18 contexts | ܣܘܪܝܐ, ܣܘܪܝܬ, ܐܣܘܪܝܐ |
+| `ܡܕܝܢ` | 1.63x | 13 contexts | ܡܕܝܢܬ, ܡܕܝܢܬܐ, ܠܡܕܝܢܬ |
+| `ܢܐܝܬ` | 1.55x | 14 contexts | ܨܝܢܐܝܬ, ܝܘܢܐܝܬ, ܝܦܢܐܝܬ |
+| `ܝܢܬܐ` | 1.71x | 9 contexts | ܩܝܢܬܐ, ܣܦܝܢܬܐ, ܡܕܝܢܬܐ |
+| `ܕܝܢܬ` | 1.67x | 9 contexts | ܡܕܝܢܬ, ܡܕܝܢܬܐ, ܠܡܕܝܢܬ |
 
 ### 6.4 Affix Compatibility (Co-occurrence)
 
@@ -467,25 +502,27 @@ Using **Recursive Hierarchical Substitutability**, we decompose complex words in
 | Word | Suggested Split | Confidence | Stem |
 |------|-----------------|------------|------|
 | ܝܘܪܕܢܢܝܬܐ | **`ܝܘܪܕܢܢ-ܝܬܐ`** | 4.5 | `ܝܘܪܕܢܢ` |
+| ܦܘܪܛܘܓܠܝܐ | **`ܦܘܪܛܘܓܠ-ܝܐ`** | 4.5 | `ܦܘܪܛܘܓܠ` |
 | ܥܘܬܡܐܢܝܬܐ | **`ܥܘܬܡܐܢ-ܝܬܐ`** | 4.5 | `ܥܘܬܡܐܢ` |
 | ܕܬܠܝܬܝܘܬܐ | **`ܕܬܠܝܬܝ-ܘܬܐ`** | 4.5 | `ܕܬܠܝܬܝ` |
 | ܕܐܢܛܝܘܟܝܐ | **`ܕܐܢܛܝܘܟ-ܝܐ`** | 4.5 | `ܕܐܢܛܝܘܟ` |
-| ܐܝܣܪܐܝܠܝܐ | **`ܐܝܣܪܐܝܠ-ܝܐ`** | 4.5 | `ܐܝܣܪܐܝܠ` |
-| ܦܘܪܛܘܓܠܝܐ | **`ܦܘܪܛܘܓܠ-ܝܐ`** | 4.5 | `ܦܘܪܛܘܓܠ` |
-| ܡܬܥܡܪܢܝܬܐ | **`ܡܬܥܡܪܢ-ܝܬܐ`** | 4.5 | `ܡܬܥܡܪܢ` |
 | ܛܘܪܥܒܕܝܢܝܐ | **`ܛܘܪܥܒܕܝܢ-ܝܐ`** | 4.5 | `ܛܘܪܥܒܕܝܢ` |
 | ܩܬܘܠܝܩܝ̈ܐ | **`ܩܬܘܠܝܩܝ-̈ܐ`** | 4.5 | `ܩܬܘܠܝܩܝ` |
+| ܡܬܥܡܪܢܝܬܐ | **`ܡܬܥܡܪܢ-ܝܬܐ`** | 4.5 | `ܡܬܥܡܪܢ` |
+| ܒܡܬܥܕܪܢܘܬܐ | **`ܒܡܬܥܕܪܢ-ܘܬܐ`** | 1.5 | `ܒܡܬܥܕܪܢ` |
+| ܬܫܥܝܬܢܝܬܐ | **`ܬܫܥܝܬܢ-ܝܬܐ`** | 1.5 | `ܬܫܥܝܬܢ` |
 | ܠܫܘܠܛܢܘܬܐ | **`ܠܫܘܠܛܢ-ܘܬܐ`** | 1.5 | `ܠܫܘܠܛܢ` |
-| ܐܘܪܬܕܘܟܣܝܐ | **`ܐܘܪܬܕܘܟܣ-ܝܐ`** | 1.5 | `ܐܘܪܬܕܘܟܣ` |
-| ܐܝܓܘܦܛܝܬܐ | **`ܐܝܓܘܦܛ-ܝܬܐ`** | 1.5 | `ܐܝܓܘܦܛ` |
-| ܘܒܐܘܚܕ̈ܢܐ | **`ܘܒܐܘܚܕ̈-ܢܐ`** | 1.5 | `ܘܒܐܘܚܕ̈` |
-| ܘܒܡܫܝܚܝܘܬܐ | **`ܘܒܡܫܝܚܝ-ܘܬܐ`** | 1.5 | `ܘܒܡܫܝܚܝ` |
-| ܢܩܪܘܡܢܛܝܐ | **`ܢܩܪܘܡܢܛ-ܝܐ`** | 1.5 | `ܢܩܪܘܡܢܛ` |
+| ܡܫܥܢܕܢܝܬܐ | **`ܡܫܥܢܕܢ-ܝܬܐ`** | 1.5 | `ܡܫܥܢܕܢ` |
+| ܐܝܣܠܢܕ̈ܝܐ | **`ܐܝܣܠܢܕ̈-ܝܐ`** | 1.5 | `ܐܝܣܠܢܕ̈` |
+| ܐܪܬܘܕܟܣܝܐ | **`ܐܪܬܘܕܟܣ-ܝܐ`** | 1.5 | `ܐܪܬܘܕܟܣ` |
+| ܦܘܠܛܝܩܝܬܐ | **`ܦܘܠܛܝܩ-ܝܬܐ`** | 1.5 | `ܦܘܠܛܝܩ` |
 
 ### 6.6 Linguistic Interpretation
 
 > **Automated Insight:**
-The language ARC appears to be more isolating or has a highly fixed vocabulary. Word-level models perform nearly as well as subword models, indicating fewer productive morphological processes.
+The language Aramaic 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
@@ -497,7 +534,7 @@ The language ARC appears to be more isolating or has a highly fixed vocabulary.
 | Component | Recommended | Rationale |
 |-----------|-------------|-----------|
 | Tokenizer | **32k BPE** | Best compression (4.58x) |
-| N-gram | **2-gram** | Lowest perplexity (365) |
+| N-gram | **2-gram** | Lowest perplexity (363) |
 | Markov | **Context-4** | Highest predictability (98.9%) |
 | Embeddings | **100d** | Balanced semantic capture and isotropy |
 
@@ -712,4 +749,4 @@ MIT License - Free for academic and commercial use.
 ---
 *Generated by Wikilangs Models Pipeline*
 
-*Report Date: 2026-01-03 05:14:47*
+*Report Date: 2026-01-03 16:33:24*

models/embeddings/aligned/arc_128d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8d628ee8950c2b7ed7c3e9de1db3782e533dcdf606048be9ab19b74823a78c1
+size 1025874179

models/embeddings/aligned/arc_128d.meta.json

@@ -0,0 +1 @@
+{"lang": "arc", "dim": 128, "max_seq_len": 512, "is_aligned": true}

models/embeddings/aligned/arc_128d.projection.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90a8bc7d91ab749fbde71ef4592ddb5930aad98cf89e4fe62609617116614d4c
+size 65664

models/embeddings/aligned/arc_128d_metadata.json

@@ -0,0 +1,8 @@
+{
+  "language": "arc",
+  "dimension": 128,
+  "version": "aligned",
+  "hub_language": "en",
+  "seed_vocab_size": 81,
+  "vocab_size": 1795
+}

models/embeddings/aligned/arc_32d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea95831f0cddf68f8d1922adb20034599b5abee861aa999d0c55c5b434433549
+size 256495619

models/embeddings/aligned/arc_32d.meta.json

@@ -0,0 +1 @@
+{"lang": "arc", "dim": 32, "max_seq_len": 512, "is_aligned": true}

models/embeddings/aligned/arc_32d.projection.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:314b43a7227e4552787faf00f8ef83eb3186cbe76d9aa61a0c07d6bc9b336111
+size 4224

models/embeddings/aligned/arc_32d_metadata.json

@@ -0,0 +1,8 @@
+{
+  "language": "arc",
+  "dimension": 32,
+  "version": "aligned",
+  "hub_language": "en",
+  "seed_vocab_size": 81,
+  "vocab_size": 1795
+}

models/embeddings/aligned/arc_64d.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80969530d1c44b57121288762e8b87141d96d2f813f5a2ce5b5770ef551329e6
+size 512955139

models/embeddings/aligned/arc_64d.meta.json

@@ -0,0 +1 @@
+{"lang": "arc", "dim": 64, "max_seq_len": 512, "is_aligned": true}

models/embeddings/aligned/arc_64d.projection.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e168e33b3341339910d3d05582308c54eda45b95a9123adf0e4b2ad948e055a
+size 16512

models/embeddings/aligned/arc_64d_metadata.json

@@ -0,0 +1,8 @@
+{
+  "language": "arc",
+  "dimension": 64,
+  "version": "aligned",
+  "hub_language": "en",
+  "seed_vocab_size": 81,
+  "vocab_size": 1795
+}

models/embeddings/monolingual/arc_128d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af294265d0648fcd88d1a3a79c52a316b5312eb2df0e22491460d7962b9abecf
-size 1025880436
+oid sha256:e8d628ee8950c2b7ed7c3e9de1db3782e533dcdf606048be9ab19b74823a78c1
+size 1025874179

models/embeddings/monolingual/arc_128d_metadata.json

@@ -11,5 +11,5 @@
     "encoding_method": "rope",
     "dim": 128
   },
-  "vocab_size": 1801
+  "vocab_size": 1795
 }

models/embeddings/monolingual/arc_32d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ca828785023acb65e4a3915ca46d6488b08273c410da1d5ebe5859bffe778de4
-size 256497268
+oid sha256:ea95831f0cddf68f8d1922adb20034599b5abee861aa999d0c55c5b434433549
+size 256495619

models/embeddings/monolingual/arc_32d_metadata.json

@@ -11,5 +11,5 @@
     "encoding_method": "rope",
     "dim": 32
   },
-  "vocab_size": 1801
+  "vocab_size": 1795
 }

models/embeddings/monolingual/arc_64d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3ca0fd83ec54d91bfbc8d6aec7c2188702f2f3ff651521ae421ae4403a931166
-size 512958324
+oid sha256:80969530d1c44b57121288762e8b87141d96d2f813f5a2ce5b5770ef551329e6
+size 512955139

models/embeddings/monolingual/arc_64d_metadata.json

@@ -11,5 +11,5 @@
     "encoding_method": "rope",
     "dim": 64
   },
-  "vocab_size": 1801
+  "vocab_size": 1795
 }

models/subword_markov/arc_markov_ctx1_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bc87b5460d601a98fbc387cbc99f283bf6cd7d98c37e6a136392f9742a5c1769
-size 61605
+oid sha256:f2cf16541ca4543598754c02901d686e966979a3f2dbef93171079a43a695800
+size 61546

models/subword_markov/arc_markov_ctx1_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 1,
   "variant": "subword",
   "language": "arc",
-  "unique_contexts": 1232,
-  "total_transitions": 368665
+  "unique_contexts": 1231,
+  "total_transitions": 367430
 }

models/subword_markov/arc_markov_ctx2_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44f7913fb0e3f66a526482ce6ecca8f16f96c74f389962eac8c930ae99d709a4
-size 236563
+oid sha256:ade6f3b86c5ac381ec52cb57cbda8cc727ed4013fd635825536cbdeb065e3851
+size 228551

models/subword_markov/arc_markov_ctx2_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 2,
   "variant": "subword",
   "language": "arc",
-  "unique_contexts": 7459,
-  "total_transitions": 367072
+  "unique_contexts": 7433,
+  "total_transitions": 365838
 }

models/subword_markov/arc_markov_ctx3_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6c02d515c8771584609106a8bfb0e1e6f4a01778f690d1e902e6c23730d85b1f
-size 614974
+oid sha256:a14f68a36dd11286c23370274ba7ff4661e8c2e5ad7150608f82b13f2e385917
+size 613765

models/subword_markov/arc_markov_ctx3_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 3,
   "variant": "subword",
   "language": "arc",
-  "unique_contexts": 28618,
-  "total_transitions": 365479
+  "unique_contexts": 28465,
+  "total_transitions": 364246
 }

models/subword_markov/arc_markov_ctx4_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ab3159c713b354d456d78d674f47cafc43e66c4c18a290510850783e3575b191
-size 1213186
+oid sha256:de9318a054b0cd1ae284c2b1e5d27ed21cc036ddfb239927d75f1c53cd6a035b
+size 1191772

models/subword_markov/arc_markov_ctx4_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 4,
   "variant": "subword",
   "language": "arc",
-  "unique_contexts": 69915,
-  "total_transitions": 363886
+  "unique_contexts": 69540,
+  "total_transitions": 362654
 }

models/subword_ngram/arc_2gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1e6f137266958bfa5bf2d5725d17daa253bc7444dd15168bf7c4aba0c82d498e
-size 30619
+oid sha256:0b0ef8b68904eb3bc261b5599c684b44c82c140473e77759ec123df2bcccf1b5
+size 30383

models/subword_ngram/arc_2gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 2,
   "variant": "subword",
   "language": "arc",
-  "unique_ngrams": 2347,
-  "total_ngrams": 368665
+  "unique_ngrams": 2330,
+  "total_ngrams": 367430
 }

models/subword_ngram/arc_3gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ce18df7b3c9474a7e74ca5ef2fd82d105d907ac676d24a02a4a055d5c06da3df
-size 135373
+oid sha256:41cf48eff4f49881b3b1ef4bee554fb58154195938b1733c73ed7e68e9e0214b
+size 134552

models/subword_ngram/arc_3gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 3,
   "variant": "subword",
   "language": "arc",
-  "unique_ngrams": 10625,
-  "total_ngrams": 367072
+  "unique_ngrams": 10583,
+  "total_ngrams": 365838
 }

models/subword_ngram/arc_4gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:36cdb55afe0905dc264f2b5348858599da9f4a343be608731c24757accd88088
-size 398536
+oid sha256:9672d51caaa89c2f58eff730631ddd563754bbaac6d78c20d8cb575cbbf65d87
+size 394701

models/subword_ngram/arc_4gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 4,
   "variant": "subword",
   "language": "arc",
-  "unique_ngrams": 28979,
-  "total_ngrams": 365479
+  "unique_ngrams": 28872,
+  "total_ngrams": 364246
 }

models/subword_ngram/arc_5gram_subword.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f92babc9e6855ff8c2b135ce81ee6f2033e836d8767833ec47cf2caf5c7f2dd
+size 537047

models/subword_ngram/arc_5gram_subword_metadata.json

@@ -0,0 +1,7 @@
+{
+  "n": 5,
+  "variant": "subword",
+  "language": "arc",
+  "unique_ngrams": 38919,
+  "total_ngrams": 362654
+}

models/tokenizer/arc_tokenizer_16k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f9d8b946b07c5ee953e8f59e73c726ea02c7e02592415beaef72b11ce24d07ed
-size 553548
+oid sha256:a79f3440c903b39d8b82904f5bdf191e3a3f69c45dc94cff5f577a68ffb722d3
+size 553519

models/tokenizer/arc_tokenizer_32k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dc878161f3ebfde37edd5364173a906b743ffa182cc4a09da6dedb402cd0d479
-size 891580
+oid sha256:f93993d85e9706c681a3124199465b9fd0ecf892ec81f742aa1dd3f2a3509aa8
+size 891227

models/tokenizer/arc_tokenizer_8k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e552460c83ce8697c66bb59c12883cdc3ee45df742f4f4a31ed47f9e096b570f
-size 390840
+oid sha256:bf1e2a73cf8ac952ea13f9fc21d2cc6279d579673c3cc68f2fc93a338ee0a5c0
+size 390696

models/vocabulary/arc_vocabulary.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a0c7a49ea4d72d65cb41e84815025c1a3ac379f03d6c2738d077d2713987df61
-size 98987
+oid sha256:6f0bb804661e7eb94ae77e3d082bdaf86902542e5dff49b802dc403e642e6d93
+size 99274

models/vocabulary/arc_vocabulary_metadata.json

@@ -1,17 +1,17 @@
 {
   "language": "arc",
-  "vocabulary_size": 6113,
+  "vocabulary_size": 6099,
   "variant": "full",
   "statistics": {
-    "type_token_ratio": 0.28987946832669004,
+    "type_token_ratio": 0.2900409450826071,
     "coverage": {
-      "top_100": 0.2557526480443379,
-      "top_1000": 0.5486970192628353,
-      "top_5000": 0.7718473583077925,
-      "top_10000": 0.8689237903161773
+      "top_100": 0.2565360778753166,
+      "top_1000": 0.5490624054041137,
+      "top_5000": 0.7721095480108975,
+      "top_10000": 0.869278442493667
     },
-    "hapax_count": 12141,
-    "hapax_ratio": 0.6651144954530513,
-    "total_documents": 1593
+    "hapax_count": 12106,
+    "hapax_ratio": 0.6649821477616039,
+    "total_documents": 1592
   }
 }

models/word_markov/arc_markov_ctx1_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:832f26dc2d4ff5a3ebd23bf31f4087617a431dd92bb988486e6d4cc21c427ab8
-size 561103
+oid sha256:5cbedb01a5447afd32845e25fc8a11d1d3d23bb575c77b402dc2373f23ba697b
+size 559495

models/word_markov/arc_markov_ctx1_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 1,
   "variant": "word",
   "language": "arc",
-  "unique_contexts": 18018,
-  "total_transitions": 61378
+  "unique_contexts": 17962,
+  "total_transitions": 61175
 }

models/word_markov/arc_markov_ctx2_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:de53c3acb316d8128ebd4e1ec3ca41812fa0d445a955fc1d817d7c545461f871
-size 1081973
+oid sha256:73c67b400e1d1d3b59ad64c463d7764d153a25bb16654c53e98cb0c5ff9eb85e
+size 1076458

models/word_markov/arc_markov_ctx2_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 2,
   "variant": "word",
   "language": "arc",
-  "unique_contexts": 45749,
-  "total_transitions": 59785
+  "unique_contexts": 45549,
+  "total_transitions": 59583
 }

models/word_markov/arc_markov_ctx3_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8d239eaf545ca57c200cb2afeabb8ab0ceb2b6cc79f4dcd2f2e5f4aa5cc4e9c1
-size 1332931
+oid sha256:5e174c8083c10c295c6f9ad804650219ba42e49b0a3d2c0566dc2c29f402d62e
+size 1328117

models/word_markov/arc_markov_ctx3_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 3,
   "variant": "word",
   "language": "arc",
-  "unique_contexts": 51822,
-  "total_transitions": 58192
+  "unique_contexts": 51591,
+  "total_transitions": 57991
 }