Upload all models and assets for ast (20251001)

README.md

@@ -23,14 +23,14 @@ dataset_info:
 metrics:
   - name: best_compression_ratio
     type: compression
-    value: 3.924
+    value: 4.427
   - name: best_isotropy
     type: isotropy
-    value: 0.7692
+    value: 0.7909
   - name: vocabulary_size
     type: vocab
-    value: 654549
-generated: 2025-12-27
+    value: 0
+generated: 2026-01-03
 ---
 
 # AST - Wikilangs Models
@@ -44,12 +44,13 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
 ### Models & Assets
 
 - Tokenizers (8k, 16k, 32k, 64k)
-- N-gram models (2, 3, 4-gram)
-- Markov chains (context of 1, 2, 3 and 4)
+- 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
+- Embeddings in various sizes and dimensions (aligned and unaligned)
 - Language Vocabulary
 - Language Statistics
+
 ![Performance Dashboard](visualizations/performance_dashboard.png)
 
 ### Analysis and Evaluation
@@ -59,7 +60,8 @@ 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. Summary & Recommendations](#6-summary--recommendations)
+- [6. Morphological Analysis (Experimental)](#6-morphological-analysis)
+- [7. Summary & Recommendations](#7-summary--recommendations)
 - [Metrics Glossary](#appendix-metrics-glossary--interpretation-guide)
 - [Visualizations Index](#visualizations-index)
 
@@ -68,71 +70,57 @@ We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and
 
 ![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** | 3.259x | 3.22 | 0.0290% | 1,033,064 |
-| **16k** | 3.531x | 3.48 | 0.0315% | 953,475 |
-| **32k** | 3.753x | 3.70 | 0.0334% | 897,137 |
-| **64k** | 3.924x 🏆 | 3.87 | 0.0350% | 858,173 |
+| **8k** | 3.569x | 3.57 | 0.0259% | 871,221 |
+| **16k** | 3.921x | 3.92 | 0.0285% | 793,006 |
+| **32k** | 4.204x | 4.21 | 0.0306% | 739,567 |
+| **64k** | 4.427x 🏆 | 4.43 | 0.0322% | 702,254 |
 
 ### Tokenization Examples
 
 Below are sample sentences tokenized with each vocabulary size:
 
-**Sample 1:** `Fechos 
-
- Personaxes importantes 
-
- Referencies 
-
- Enllaces esternos 
-
-Categoría...`
+**Sample 1:** `Luiz Diallisson de Souza Alves ye un futbolista brasilanu. Clubes Kuban Referenc...`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 8k | `▁fechos ▁personaxes ▁importantes ▁referencies ▁enllaces ▁esternos ▁categoría : sieglu ▁viii ... (+4 more)` | 14 |
-| 16k | `▁fechos ▁personaxes ▁importantes ▁referencies ▁enllaces ▁esternos ▁categoría : sieglu ▁viii ... (+4 more)` | 14 |
-| 32k | `▁fechos ▁personaxes ▁importantes ▁referencies ▁enllaces ▁esternos ▁categoría : sieglu ▁viii ... (+4 more)` | 14 |
-| 64k | `▁fechos ▁personaxes ▁importantes ▁referencies ▁enllaces ▁esternos ▁categoría : sieglu ▁viii ... (+4 more)` | 14 |
+| 8k | `▁lu iz ▁di all is son ▁de ▁sou za ▁al ... (+14 more)` | 24 |
+| 16k | `▁lu iz ▁di all is son ▁de ▁sou za ▁al ... (+14 more)` | 24 |
+| 32k | `▁luiz ▁di all is son ▁de ▁souza ▁alves ▁ye ▁un ... (+11 more)` | 21 |
+| 64k | `▁luiz ▁di all isson ▁de ▁souza ▁alves ▁ye ▁un ▁futbolista ... (+10 more)` | 20 |
 
-**Sample 2:** `Armental ye un llugar de la parroquia de Talarén nel conceyu asturianu de Navia....`
+**Sample 2:** `Vagner da Silva Sarti ye un ex-futbolista brasilanu. Clubes Referencies Enllaces...`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 8k | `▁ar mental ▁ye ▁un ▁llugar ▁de ▁la ▁parroquia ▁de ▁tal ... (+18 more)` | 28 |
-| 16k | `▁ar mental ▁ye ▁un ▁llugar ▁de ▁la ▁parroquia ▁de ▁tal ... (+18 more)` | 28 |
-| 32k | `▁ar mental ▁ye ▁un ▁llugar ▁de ▁la ▁parroquia ▁de ▁tal ... (+16 more)` | 26 |
-| 64k | `▁ar mental ▁ye ▁un ▁llugar ▁de ▁la ▁parroquia ▁de ▁tal ... (+16 more)` | 26 |
-
-**Sample 3:** `Fechos 
-  -
-
- Nacencies 
-  -
-
- Muertes 
-  -
-
- Referencies 
+| 8k | `▁va gn er ▁da ▁silva ▁sar ti ▁ye ▁un ▁ex ... (+10 more)` | 20 |
+| 16k | `▁va gner ▁da ▁silva ▁sar ti ▁ye ▁un ▁ex - ... (+9 more)` | 19 |
+| 32k | `▁va gner ▁da ▁silva ▁sar ti ▁ye ▁un ▁ex - ... (+9 more)` | 19 |
+| 64k | `▁va gner ▁da ▁silva ▁sar ti ▁ye ▁un ▁ex - ... (+9 more)` | 19 |
 
- Enllaces esternos 
-...`
+**Sample 3:** `(MMLXXXII) va ser un añu normal entamáu en xueves nel calendariu gregorianu. Ref...`
 
 | Vocab | Tokens | Count |
 |-------|--------|-------|
-| 8k | `▁fechos ▁- ▁nacencies ▁- ▁muertes ▁- ▁referencies ▁enllaces ▁esternos ▁categoría ... (+7 more)` | 17 |
-| 16k | `▁fechos ▁- ▁nacencies ▁- ▁muertes ▁- ▁referencies ▁enllaces ▁esternos ▁categoría ... (+7 more)` | 17 |
-| 32k | `▁fechos ▁- ▁nacencies ▁- ▁muertes ▁- ▁referencies ▁enllaces ▁esternos ▁categoría ... (+7 more)` | 17 |
-| 64k | `▁fechos ▁- ▁nacencies ▁- ▁muertes ▁- ▁referencies ▁enllaces ▁esternos ▁categoría ... (+7 more)` | 17 |
+| 8k | `▁( m m l xx x ii ) ▁va ▁ser ... (+17 more)` | 27 |
+| 16k | `▁( mm l xx x ii ) ▁va ▁ser ▁un ... (+14 more)` | 24 |
+| 32k | `▁( mm l xx xii ) ▁va ▁ser ▁un ▁añu ... (+12 more)` | 22 |
+| 64k | `▁( mm lxx xii ) ▁va ▁ser ▁un ▁añu ▁normal ... (+11 more)` | 21 |
 
 
 ### Key Findings
 
-- **Best Compression:** 64k achieves 3.924x compression
-- **Lowest UNK Rate:** 8k with 0.0290% unknown tokens
+- **Best Compression:** 64k achieves 4.427x compression
+- **Lowest UNK Rate:** 8k with 0.0259% unknown tokens
 - **Trade-off:** Larger vocabularies improve compression but increase model size
 - **Recommendation:** 32k vocabulary provides optimal balance for production use
 
@@ -141,57 +129,89 @@ Categoría...`
 
 ![N-gram Perplexity](visualizations/ngram_perplexity.png)
 
+![N-gram Unique](visualizations/ngram_unique.png)
+
 ![N-gram Coverage](visualizations/ngram_coverage.png)
 
 ### Results
 
-| N-gram | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
-|--------|------------|---------|----------------|------------------|-------------------|
-| **2-gram** | 95,540 🏆 | 16.54 | 1,568,799 | 13.6% | 26.9% |
-| **2-gram** | 311 🏆 | 8.28 | 23,389 | 65.5% | 98.4% |
-| **3-gram** | 573,984 | 19.13 | 3,974,147 | 5.1% | 13.4% |
-| **3-gram** | 2,766 | 11.43 | 195,082 | 25.8% | 68.5% |
-| **4-gram** | 1,609,317 | 20.62 | 7,247,181 | 3.9% | 9.3% |
-| **4-gram** | 16,954 | 14.05 | 1,178,742 | 12.7% | 37.0% |
+| N-gram | Variant | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
+|--------|---------|------------|---------|----------------|------------------|-------------------|
+| **2-gram** | Word | 133,027 | 17.02 | 1,354,323 | 9.8% | 21.6% |
+| **2-gram** | Subword | 260 🏆 | 8.02 | 19,069 | 69.7% | 99.1% |
+| **3-gram** | Word | 646,899 | 19.30 | 2,908,394 | 4.2% | 10.7% |
+| **3-gram** | Subword | 2,223 | 11.12 | 139,212 | 28.0% | 72.3% |
+| **4-gram** | Word | 1,559,764 | 20.57 | 4,707,856 | 3.3% | 7.5% |
+| **4-gram** | Subword | 13,372 | 13.71 | 791,795 | 13.9% | 39.3% |
 
 ### Top 5 N-grams by Size
 
-**2-grams:**
+**2-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `de la` | 891,402 |
+| 2 | `de los` | 329,410 |
+| 3 | `la so` | 220,083 |
+| 4 | `a la` | 215,036 |
+| 5 | `de les` | 208,071 |
+
+**3-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `referencies enllaces esternos` | 101,643 |
+| 2 | `de la so` | 48,838 |
+| 3 | `d estaos xuníos` | 34,333 |
+| 4 | `enllaces esternos de` | 33,237 |
+| 5 | `una población de` | 30,269 |
+
+**4-grams (Word):**
+
+| Rank | N-gram | Count |
+|------|--------|-------|
+| 1 | `referencies enllaces esternos de` | 32,314 |
+| 2 | `tien una población de` | 26,720 |
+| 3 | `una población de y` | 19,598 |
+| 4 | `y una superficie de` | 19,554 |
+| 5 | `una superficie de km` | 19,519 |
+
+**2-grams (Subword):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `d '` | 1,196,313 |
-| 2 | `de la` | 875,667 |
-| 3 | `' l` | 534,478 |
-| 4 | `| |` | 438,858 |
-| 5 | `l '` | 403,691 |
+| 1 | `a _` | 12,346,491 |
+| 2 | `e _` | 10,275,492 |
+| 3 | `s _` | 10,054,248 |
+| 4 | `_ d` | 9,863,919 |
+| 5 | `e s` | 9,411,923 |
 
-**3-grams:**
+**3-grams (Subword):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `| - |` | 128,285 |
-| 2 | `referencies enllaces esternos` | 104,162 |
-| 3 | `- | |` | 89,758 |
-| 4 | `- - -` | 81,514 |
-| 5 | `d ' un` | 69,529 |
+| 1 | `_ d e` | 7,215,701 |
+| 2 | `d e _` | 5,349,035 |
+| 3 | `e s _` | 4,769,369 |
+| 4 | `o s _` | 3,909,790 |
+| 5 | `l a _` | 3,068,189 |
 
-**4-grams:**
+**4-grams (Subword):**
 
 | Rank | N-gram | Count |
 |------|--------|-------|
-| 1 | `- - - -` | 69,470 |
-| 2 | `enllaces esternos categoría :` | 63,833 |
-| 3 | `referencies enllaces esternos categoría` | 60,665 |
-| 4 | `. referencies enllaces esternos` | 51,144 |
-| 5 | `| linear | -` | 50,481 |
+| 1 | `_ d e _` | 4,975,922 |
+| 2 | `_ l a _` | 2,468,941 |
+| 3 | `d e _ l` | 1,667,072 |
+| 4 | `a _ d e` | 1,422,241 |
+| 5 | `s _ d e` | 1,380,334 |
 
 
 ### Key Findings
 
-- **Best Perplexity:** 2-gram with 311
+- **Best Perplexity:** 2-gram (subword) with 260
 - **Entropy Trend:** Decreases with larger n-grams (more predictable)
-- **Coverage:** Top-1000 patterns cover ~37% of corpus
+- **Coverage:** Top-1000 patterns cover ~39% of corpus
 - **Recommendation:** 4-gram or 5-gram for best predictive performance
 
 ---
@@ -199,55 +219,86 @@ Categoría...`
 
 ![Markov Entropy](visualizations/markov_entropy.png)
 
+![Markov Contexts](visualizations/markov_contexts.png)
+
 ![Markov Branching](visualizations/markov_branching.png)
 
 ### Results
 
-| Context | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
-|---------|-------------|------------|------------------|-----------------|----------------|
-| **1** | 0.7150 | 1.641 | 8.69 | 1,669,949 | 28.5% |
-| **1** | 1.5193 | 2.866 | 10.68 | 8,875 | 0.0% |
-| **2** | 0.4611 | 1.377 | 2.90 | 14,499,551 | 53.9% |
-| **2** | 0.7271 | 1.655 | 4.90 | 94,766 | 27.3% |
-| **3** | 0.2234 | 1.167 | 1.58 | 42,031,886 | 77.7% |
-| **3** | 0.8068 | 1.749 | 4.59 | 464,259 | 19.3% |
-| **4** | 0.1062 🏆 | 1.076 | 1.22 | 66,322,442 | 89.4% |
-| **4** | 0.7182 🏆 | 1.645 | 3.49 | 2,131,889 | 28.2% |
+| Context | Variant | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
+|---------|---------|-------------|------------|------------------|-----------------|----------------|
+| **1** | Word | 1.0381 | 2.054 | 12.99 | 1,204,316 | 0.0% |
+| **1** | Subword | 1.1983 | 2.295 | 7.97 | 10,463 | 0.0% |
+| **2** | Word | 0.4193 | 1.337 | 2.57 | 15,634,564 | 58.1% |
+| **2** | Subword | 0.6558 | 1.576 | 4.28 | 83,437 | 34.4% |
+| **3** | Word | 0.1865 | 1.138 | 1.44 | 40,202,890 | 81.4% |
+| **3** | Subword | 0.6846 | 1.607 | 4.03 | 357,207 | 31.5% |
+| **4** | Word | 0.0789 🏆 | 1.056 | 1.15 | 57,817,277 | 92.1% |
+| **4** | Subword | 0.6846 | 1.607 | 3.51 | 1,439,883 | 31.5% |
+
+### Generated Text Samples (Word-based)
+
+Below are text samples generated from each word-based Markov chain model:
+
+**Context Size 1:**
+
+1. `de xunu empieza a centros multimodales funciones nel nectariu semilunar o n ucraín y comunicaciones ...`
+2. `la pieza cornelius coffin fit cuando el algebasó 503 mariña d una solución bonal o nun`
+3. `y 15 m sobre l uniforme del postreru gran midida china tales from here mirror weekly`
+
+**Context Size 2:**
+
+1. `de la provincia dende esti tornéu surdió en y persuadió a eliza dushku en películes d estudiante`
+2. `de los documentos relativos al mercáu l so antiguu nome dau más tarde l empresariu estremeñu dueñu`
+3. `la so base na isla parker llogró atrapar la pelota vasca que se llevó a empecipiar una`
+
+**Context Size 3:**
+
+1. `referencies enllaces esternos el salín nel suelu y ente vexetación trupa sicasí en marismas y ribere...`
+2. `de la so agua h havagazı gas o otobüs bus y t troleybüs trolebús magar que los entamos`
+3. `enllaces esternos de côte d or na rexón de gran este llenda con tien una población de 1`
+
+**Context Size 4:**
+
+1. `referencies enllaces esternos de saboya de francia de bretaña de dreux de bretaña`
+2. `tien una población de 1 690 471 habitantes y un puertu fluvial sobre l paraná amás tien importancia ...`
+3. `una población de y una superficie de km ver tamién referencies enllaces esternos de xapón de la pref...`
+
 
-### Generated Text Samples
+### Generated Text Samples (Subword-based)
 
-Below are text samples generated from each Markov chain model:
+Below are text samples generated from each subword-based Markov chain model:
 
 **Context Size 1:**
 
-1. `de gossip girl play ye como xenofonte en dussel , y el 7 d ' amuesa`
-2. `, pero la botánica referencies ver , collaboró en valdivia . mientres la humanidá al chinu`
-3. `. ( 2 ) - ḥḏ horusmuriu blancumennefermenfismit rahina . isbn 0 british lion , yera`
+1. `_fén_an_yaconyíc`
+2. `er,_ciesunton_a_`
+3. `a_tostelociz_ce_`
 
 **Context Size 2:**
 
-1. `d ' ellos yera detectáu polos enemigos . shiva prakash ( 1997 ) , nel conceyu sevillanu`
-2. `de la litografía y l ' ala posterior : chronica majora : una « inocente ya inconsciente`
-3. `' l xeneral prats tamién pudo ante fernando verdasco david ferrer por 6 - 2 | ríu`
+1. `a_gasainel_tabaro`
+2. `e_es_de_y_chel_má`
+3. `s_astamudia_de_ll`
 
 **Context Size 3:**
 
-1. `| - | 38378 - | | 1997 tb18 | | 4 | align = right | [`
-2. `referencies enllaces esternos categoría : montserrat`
-3. `- | | 2001 sd35 | | 16 | | 592 | | < small > 1911 <`
+1. `_de_scharacióse_le`
+2. `de_tragar_primera_`
+3. `es_so_títulu_miliz`
 
 **Context Size 4:**
 
-1. `- - - - - - - - - - - - - - - - - - -`
-2. `enllaces esternos categoría : pintores de parís categoría : sabios de la torre eiffel , los nacional...`
-3. `referencies enllaces esternos categoría : comuñes de nord`
+1. `_de_los_sobres_del_`
+2. `_la_ermistoria_dife`
+3. `de_los_xeneia,_cons`
 
 
 ### Key Findings
 
-- **Best Predictability:** Context-4 with 89.4% predictability
+- **Best Predictability:** Context-4 (word) with 92.1% predictability
 - **Branching Factor:** Decreases with context size (more deterministic)
-- **Memory Trade-off:** Larger contexts require more storage (2,131,889 contexts)
+- **Memory Trade-off:** Larger contexts require more storage (1,439,883 contexts)
 - **Recommendation:** Context-3 or Context-4 for text generation
 
 ---
@@ -263,64 +314,64 @@ Below are text samples generated from each Markov chain model:
 
 | Metric | Value |
 |--------|-------|
-| Vocabulary Size | 654,549 |
-| Total Tokens | 80,184,102 |
-| Mean Frequency | 122.50 |
+| Vocabulary Size | 555,056 |
+| Total Tokens | 75,071,637 |
+| Mean Frequency | 135.25 |
 | Median Frequency | 4 |
-| Frequency Std Dev | 8722.95 |
+| Frequency Std Dev | 9337.66 |
 
 ### Most Common Words
 
 | Rank | Word | Frequency |
 |------|------|-----------|
-| 1 | de | 5,075,921 |
-| 2 | la | 2,521,840 |
-| 3 | y | 2,071,360 |
-| 4 | d | 1,229,266 |
-| 5 | a | 1,176,335 |
-| 6 | del | 1,090,980 |
-| 7 | en | 1,071,173 |
-| 8 | que | 1,020,518 |
-| 9 | los | 971,499 |
-| 10 | l | 968,352 |
+| 1 | de | 4,994,843 |
+| 2 | la | 2,512,518 |
+| 3 | y | 2,055,358 |
+| 4 | d | 1,181,646 |
+| 5 | a | 1,163,388 |
+| 6 | del | 1,091,464 |
+| 7 | en | 1,070,328 |
+| 8 | que | 1,013,684 |
+| 9 | los | 966,280 |
+| 10 | l | 958,680 |
 
 ### Least Common Words (from vocabulary)
 
 | Rank | Word | Frequency |
 |------|------|-----------|
-| 1 | leptafeke | 2 |
-| 2 | haua | 2 |
-| 3 | küzdoblani | 2 |
-| 4 | contrarrellatu | 2 |
-| 5 | semilleru | 2 |
-| 6 | bisterca | 2 |
-| 7 | šafarsko | 2 |
-| 8 | vyfalu | 2 |
-| 9 | ribich | 2 |
-| 10 | lacos | 2 |
+| 1 | sverlo | 2 |
+| 2 | kmca | 2 |
+| 3 | antimaterialistas | 2 |
+| 4 | infectados | 2 |
+| 5 | historietistas | 2 |
+| 6 | curtmetratxe | 2 |
+| 7 | rugna | 2 |
+| 8 | lleáu | 2 |
+| 9 | queña | 2 |
+| 10 | nkoghe | 2 |
 
 ### Zipf's Law Analysis
 
 | Metric | Value |
 |--------|-------|
-| Zipf Coefficient | 1.0077 |
-| R² (Goodness of Fit) | 0.995140 |
+| Zipf Coefficient | 0.9991 |
+| R² (Goodness of Fit) | 0.995555 |
 | Adherence Quality | **excellent** |
 
 ### Coverage Analysis
 
 | Top N Words | Coverage |
 |-------------|----------|
-| Top 100 | 40.0% |
-| Top 1,000 | 60.0% |
-| Top 5,000 | 76.4% |
-| Top 10,000 | 82.7% |
+| Top 100 | 41.7% |
+| Top 1,000 | 60.8% |
+| Top 5,000 | 76.8% |
+| Top 10,000 | 83.1% |
 
 ### Key Findings
 
-- **Zipf Compliance:** R²=0.9951 indicates excellent adherence to Zipf's law
-- **High Frequency Dominance:** Top 100 words cover 40.0% of corpus
-- **Long Tail:** 644,549 words needed for remaining 17.3% coverage
+- **Zipf Compliance:** R²=0.9956 indicates excellent adherence to Zipf's law
+- **High Frequency Dominance:** Top 100 words cover 41.7% of corpus
+- **Long Tail:** 545,056 words needed for remaining 16.9% coverage
 
 ---
 ## 5. Word Embeddings Evaluation
@@ -333,24 +384,130 @@ Below are text samples generated from each Markov chain model:
 
 ![t-SNE Sentences](visualizations/tsne_sentences.png)
 
-### Model Comparison
 
-| Model | Vocab Size | Dimension | Avg Norm | Std Norm | Isotropy |
-|-------|------------|-----------|----------|----------|----------|
-| **mono_32d** | 510,373 | 32 | 3.008 | 0.935 | 0.7692 🏆 |
-| **mono_64d** | 510,373 | 64 | 3.395 | 0.938 | 0.7616 |
-| **mono_128d** | 510,373 | 128 | 3.842 | 0.965 | 0.6988 |
-| **embeddings_enhanced** | 0 | 0 | 0.000 | 0.000 | 0.0000 |
+### 5.1 Cross-Lingual Alignment
+
+> *Note: Multilingual alignment visualization not available for this language.*
+
+
+### 5.2 Model Comparison
+
+| Model | Dimension | Isotropy | Semantic Density | Alignment R@1 | Alignment R@10 |
+|-------|-----------|----------|------------------|---------------|----------------|
+| **mono_32d** | 32 | 0.7909 🏆 | 0.3827 | N/A | N/A |
+| **mono_64d** | 64 | 0.7802 | 0.3065 | N/A | N/A |
+| **mono_128d** | 128 | 0.7192 | 0.2391 | N/A | N/A |
 
 ### Key Findings
 
-- **Best Isotropy:** mono_32d with 0.7692 (more uniform distribution)
-- **Dimension Trade-off:** Higher dimensions capture more semantics but reduce isotropy
-- **Vocabulary Coverage:** All models cover 510,373 words
-- **Recommendation:** 100d for balanced semantic capture and efficiency
+- **Best Isotropy:** mono_32d with 0.7909 (more uniform distribution)
+- **Semantic Density:** Average pairwise similarity of 0.3094. Lower values indicate better semantic separation.
+- **Alignment Quality:** No aligned models evaluated in this run.
+- **Recommendation:** 128d aligned for best cross-lingual performance
 
 ---
-## 6. Summary & Recommendations
+## 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 | - |
+
+### 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 |
+|--------|----------|
+| `-co` | comíen, compelidos, conciliable |
+| `-ma` | maravíase, maça, matematización |
+| `-re` | reescalada, reprimió, reconociéralu |
+| `-de` | deduz, declaratorio, desfila |
+| `-ca` | caminómetru, castromil, caecilia |
+
+#### Productive Suffixes
+| Suffix | Examples |
+|--------|----------|
+| `-s` | phrygilus, anticolinérgicos, friulianos |
+| `-a` | raksasa, estendería, reescalada |
+| `-es` | ibes, distopíes, ziríes |
+| `-os` | anticolinérgicos, friulianos, afogadiegos |
+| `-se` | esmoreciérase, maravíase, cuayábase |
+| `-as` | monarquistas, gorgas, mimeografiadas |
+| `-en` | altshausen, comíen, blegen |
+
+### 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 |
+|------|----------|------------------|----------|
+| `iend` | 1.80x | 206 contexts | fiend, iendo, viendi |
+| `renc` | 2.05x | 99 contexts | frenc, wrench, rencor |
+| `ient` | 1.67x | 271 contexts | vient, iente, aient |
+| `enci` | 1.52x | 262 contexts | venci, benci, cenci |
+| `acio` | 1.63x | 166 contexts | nacio, cacio, tacio |
+| `ació` | 1.79x | 94 contexts | lació, xació, ñació |
+| `nter` | 1.38x | 335 contexts | inter, enter, unter |
+| `ontr` | 1.63x | 118 contexts | contr, contra, montra |
+| `ener` | 1.42x | 205 contexts | enerc, tener, enero |
+| `ntos` | 1.79x | 67 contexts | antos, entos, tintos |
+| `ntes` | 1.49x | 144 contexts | antes, entes, fontes |
+| `efer` | 1.61x | 86 contexts | sefer, nefer, refer |
+
+### 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 |
+|--------|--------|-----------|----------|
+| `-co` | `-s` | 59 words | conversas, concinnus |
+| `-ca` | `-s` | 53 words | cancelaciones, caloiros |
+| `-ca` | `-a` | 49 words | cartajima, campana |
+| `-co` | `-a` | 44 words | comella, copia |
+| `-ma` | `-a` | 38 words | matina, matrioshka |
+| `-re` | `-s` | 34 words | rectos, restaurantes |
+| `-ma` | `-s` | 31 words | maniobres, maderensis |
+| `-de` | `-s` | 31 words | descatados, definitives |
+| `-co` | `-es` | 25 words | cotidales, coleicionables |
+| `-re` | `-a` | 24 words | renombraría, retomara |
+
+### 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 |
+|------|-----------------|------------|------|
+| retractores | **`re-tractor-es`** | 6.0 | `tractor` |
+| aseguráronse | **`aseguráron-se`** | 4.5 | `aseguráron` |
+| tendiéronse | **`tendiéron-se`** | 4.5 | `tendiéron` |
+| tresversales | **`tresversal-es`** | 4.5 | `tresversal` |
+| redefiniéronse | **`re-de-finiéron-se`** | 4.5 | `finiéron` |
+| redistributivo | **`re-distributivo`** | 4.5 | `distributivo` |
+| prométese | **`prométe-se`** | 4.5 | `prométe` |
+| escaecíen | **`escaecí-en`** | 4.5 | `escaecí` |
+| domadores | **`domador-es`** | 4.5 | `domador` |
+| consérvense | **`co-nsérv-en-se`** | 4.5 | `nsérv` |
+| descripto | **`de-scripto`** | 4.5 | `scripto` |
+| esaxeróse | **`esaxeró-se`** | 4.5 | `esaxeró` |
+| acentores | **`acentor-es`** | 4.5 | `acentor` |
+| detrayendo | **`de-trayendo`** | 4.5 | `trayendo` |
+| renormalización | **`re-normalización`** | 4.5 | `normalización` |
+
+### 6.6 Linguistic Interpretation
+
+> **Automated Insight:**
+The language AST 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.
+
+---
+## 7. Summary & Recommendations
 
 ![Performance Dashboard](visualizations/performance_dashboard.png)
 
@@ -358,11 +515,12 @@ Below are text samples generated from each Markov chain model:
 
 | Component | Recommended | Rationale |
 |-----------|-------------|-----------|
-| Tokenizer | **32k BPE** | Best compression (3.92x) with low UNK rate |
-| N-gram | **5-gram** | Lowest perplexity (311) |
-| Markov | **Context-4** | Highest predictability (89.4%) |
+| Tokenizer | **64k BPE** | Best compression (4.43x) |
+| N-gram | **2-gram** | Lowest perplexity (260) |
+| Markov | **Context-4** | Highest predictability (92.1%) |
 | Embeddings | **100d** | Balanced semantic capture and isotropy |
 
+
 ---
 ## Appendix: Metrics Glossary & Interpretation Guide
 
@@ -552,7 +710,8 @@ If you use these models in your research, please cite:
   author = {Kamali, Omar},
   title = {Wikilangs: Open NLP Models for Wikipedia Languages},
   year = {2025},
-  publisher = {HuggingFace},
+  doi = {10.5281/zenodo.18073153},
+  publisher = {Zenodo},
   url = {https://huggingface.co/wikilangs}
   institution = {Omneity Labs}
 }
@@ -568,7 +727,8 @@ MIT License - Free for academic and commercial use.
 - 🤗 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: 2025-12-27 20:35:27*
+*Report Date: 2026-01-03 09:38:21*

models/embeddings/monolingual/ast_128d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:39f5f134a23b0ae5b3bba7a4864ecc3599e28dd6a5c44748c538e6eb50542a1b
-size 1555994683
+oid sha256:8c4500458da720e9717e0be530291de837dfe1937d937d8f59cd5975214f6d01
+size 1526234273

models/embeddings/monolingual/ast_128d_metadata.json

@@ -3,11 +3,13 @@
   "dimension": 128,
   "version": "monolingual",
   "training_params": {
-    "dim": 128,
+    "algorithm": "skipgram",
     "min_count": 5,
     "window": 5,
     "negative": 5,
-    "epochs": 5
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 128
   },
-  "vocab_size": 510373
+  "vocab_size": 481846
 }

models/embeddings/monolingual/ast_32d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:aecd461d0dd299f160b9fefcb4d6bbc47ffe601253410b6bf66aa59353683649
-size 396028219
+oid sha256:aa3cdba099d3af49ab89043ad5713c1fe468060d520334f5200b000925a5e30e
+size 388176545

models/embeddings/monolingual/ast_32d_metadata.json

@@ -3,11 +3,13 @@
   "dimension": 32,
   "version": "monolingual",
   "training_params": {
-    "dim": 32,
+    "algorithm": "skipgram",
     "min_count": 5,
     "window": 5,
     "negative": 5,
-    "epochs": 5
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 32
   },
-  "vocab_size": 510373
+  "vocab_size": 481846
 }

models/embeddings/monolingual/ast_64d.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a7552caa3e9f8ef8607c05afb5b9266e4c9a3e70fdba272075a7b9912e995045
-size 782683707
+oid sha256:7e24d4561e7c9f15e391c915d7d4096c8b8adf794324cd6df9a60e78e6f5fc6a
+size 767529121

models/embeddings/monolingual/ast_64d_metadata.json

@@ -3,11 +3,13 @@
   "dimension": 64,
   "version": "monolingual",
   "training_params": {
-    "dim": 64,
+    "algorithm": "skipgram",
     "min_count": 5,
     "window": 5,
     "negative": 5,
-    "epochs": 5
+    "epochs": 5,
+    "encoding_method": "rope",
+    "dim": 64
   },
-  "vocab_size": 510373
+  "vocab_size": 481846
 }

models/subword_markov/ast_markov_ctx1_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:797c1dfcb6c516214d2dc02ed941e64e840e49d291e2cc758d5482d86ae04640
-size 602821
+oid sha256:1b9be5fbc395f6cb378f05193e1ba7234945d9bca4e71b4cdad6aea648215809
+size 555949

models/subword_markov/ast_markov_ctx1_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 1,
   "variant": "subword",
   "language": "ast",
-  "unique_contexts": 8875,
-  "total_transitions": 508763835
+  "unique_contexts": 10463,
+  "total_transitions": 464881103
 }

models/subword_markov/ast_markov_ctx2_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e0ee540fb4354bf9779b55ea96cde4252defd9f562e14d445ac6b6c9b457c3f1
-size 3577755
+oid sha256:a0a7c5f99ccca8b527cebd7afb3189b4a99a6cc219cd1b861b501d17180c93b0
+size 2891072

models/subword_markov/ast_markov_ctx2_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 2,
   "variant": "subword",
   "language": "ast",
-  "unique_contexts": 94766,
-  "total_transitions": 508625606
+  "unique_contexts": 83437,
+  "total_transitions": 464744539
 }

models/subword_markov/ast_markov_ctx3_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:db3f45e2155b76125f075309cbec6b285d5ee2401fd063d4b596638cb534193e
-size 16073938
+oid sha256:6c060d1174629162cb8911f78054119018ce6e2fc628936bfdeb36691d35489c
+size 11949355

models/subword_markov/ast_markov_ctx3_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 3,
   "variant": "subword",
   "language": "ast",
-  "unique_contexts": 464259,
-  "total_transitions": 508487377
+  "unique_contexts": 357207,
+  "total_transitions": 464607975
 }

models/subword_markov/ast_markov_ctx4_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dd0b9d4e9db8c19eeb444ba13e49b7a48ec4b84512f949ef315285aff0822a0e
-size 57292008
+oid sha256:60988880010bce63ffacea78379f2d4b447c453dc44f6445a08b1d6e123f645c
+size 39925220

models/subword_markov/ast_markov_ctx4_subword_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 4,
   "variant": "subword",
   "language": "ast",
-  "unique_contexts": 2131889,
-  "total_transitions": 508349148
+  "unique_contexts": 1439883,
+  "total_transitions": 464471411
 }

models/subword_ngram/ast_2gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:417e9408d0c904ed21f23501e55082a7494276c2cfce4650855bdde090aa9e15
-size 314961
+oid sha256:0c50ba055002320ed277521dcbb25feecc7139a7b02e23e72ca8255ada347410
+size 259482

models/subword_ngram/ast_2gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 2,
   "variant": "subword",
   "language": "ast",
-  "unique_ngrams": 23389,
-  "total_ngrams": 508763835
+  "unique_ngrams": 19069,
+  "total_ngrams": 464881103
 }

models/subword_ngram/ast_3gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:730e803a7b1373e12874f3e3ee787c6229e54c4f714b7b5b45ecbc90e7e1f2ef
-size 2302607
+oid sha256:d1ee7e63a9f92e3bf668df5f64f593ca6671568eb54bc02ab9e3b50ee2ae7ae4
+size 1721257

models/subword_ngram/ast_3gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 3,
   "variant": "subword",
   "language": "ast",
-  "unique_ngrams": 195082,
-  "total_ngrams": 508625606
+  "unique_ngrams": 139212,
+  "total_ngrams": 464744539
 }

models/subword_ngram/ast_4gram_subword.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b418657cc08b4f91e0b2c2eda00aef5c4302d754fe9dccc5feb84ff1b24a6291
-size 13600667
+oid sha256:390181c0c7fcc8b118b108e18cb830493109f584b4b87f76850993e3e9dda0be
+size 9353492

models/subword_ngram/ast_4gram_subword_metadata.json

@@ -2,6 +2,6 @@
   "n": 4,
   "variant": "subword",
   "language": "ast",
-  "unique_ngrams": 1178742,
-  "total_ngrams": 508487377
+  "unique_ngrams": 791795,
+  "total_ngrams": 464607975
 }

models/tokenizer/ast_tokenizer_16k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:907f4bece802b1d3656fdf65b960935cdd809ca5733b7b73b0823ce18af2c113
-size 513109
+oid sha256:f5014581384d443a91ffa0f71150e3ecc09244a1056ad046785f8b8cbc4ce42c
+size 511145

models/tokenizer/ast_tokenizer_32k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:88d4563db9be715e9225ff77615f89d344e1f2db1b32e1119ffcb1adf873068a
-size 795952
+oid sha256:13c458e53178d158dee7ca9d93a562de7bf19975484f7e21af497697c021bb6a
+size 791500

models/tokenizer/ast_tokenizer_64k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bab637cbda657ea59afedbac3a722b8a7d18839ff031e3a67f1d79f914cb1edd
-size 1371599
+oid sha256:b619430c9173d8e202e7905a328c5bca3d915f85a3540adb29b3d6230f10d749
+size 1364026

models/tokenizer/ast_tokenizer_8k.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:53538ed204ba1e90b9a3ec72f89f3a9639ddf5f7e06a5d91daf2182e1dbacf32
-size 374583
+oid sha256:8172313a7a1d41e2c61618ad0fab8a1cbd16688f798c007cba80fce5c693133c
+size 374227

models/vocabulary/ast_vocabulary.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e95a025d0a2e0d5c5fc9e67b2724543af5eaadf84e83944b5601b9e113deee95
-size 9493041
+oid sha256:42abd03dfba1393adae66a9e07842fcb85830c5f2dbcb0e037ab56b27fbc8846
+size 8277795

models/vocabulary/ast_vocabulary_metadata.json

@@ -1,16 +1,17 @@
 {
   "language": "ast",
-  "vocabulary_size": 654549,
+  "vocabulary_size": 555056,
+  "variant": "full",
   "statistics": {
-    "type_token_ratio": 0.020582308161438793,
+    "type_token_ratio": 0.015923489955309757,
     "coverage": {
-      "top_100": 0.39513983436676664,
-      "top_1000": 0.5923739422740554,
-      "top_5000": 0.7541024014353347,
-      "top_10000": 0.8168048946874377
+      "top_100": 0.4129270164572954,
+      "top_1000": 0.6027767233040656,
+      "top_5000": 0.7612431178669916,
+      "top_10000": 0.8237422388332004
     },
-    "hapax_count": 1016752,
-    "hapax_ratio": 0.608359595309283,
-    "total_documents": 138229
+    "hapax_count": 650708,
+    "hapax_ratio": 0.5396644782892838,
+    "total_documents": 136564
   }
 }

models/word_markov/ast_markov_ctx1_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e50511b1eedd319fddb7babd82679e20996cdc8c40bb7414b7729fa744061d38
-size 125168928
+oid sha256:df2720891f77ec4fc87c424807076f856c9ebed7d35e8609e740c581e849ba43
+size 130011519

models/word_markov/ast_markov_ctx1_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 1,
   "variant": "word",
   "language": "ast",
-  "unique_contexts": 1669949,
-  "total_transitions": 101266716
+  "unique_contexts": 1204316,
+  "total_transitions": 75585781
 }

models/word_markov/ast_markov_ctx2_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5e97c5fda98a744b9ecdbc923910e1c4de28e3c699d3c7a3519ea2d5826e515a
-size 436569711
+oid sha256:360e77ba671d9d79ef0176080b3098535eb366bbb657d0a5d965de6cbcaaf01a
+size 443953571

models/word_markov/ast_markov_ctx2_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 2,
   "variant": "word",
   "language": "ast",
-  "unique_contexts": 14499551,
-  "total_transitions": 101128488
+  "unique_contexts": 15634564,
+  "total_transitions": 75449217
 }

models/word_markov/ast_markov_ctx3_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3d70d8996da08da859f83f17208f24ec5fe8a079a8cdc33b458965bdf29e6a65
-size 863814981
+oid sha256:dd3378d79db7d6144445a3f22f70dd19d9f2db6b768c7e3368bcecdeffa20cbe
+size 807770212

models/word_markov/ast_markov_ctx3_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 3,
   "variant": "word",
   "language": "ast",
-  "unique_contexts": 42031886,
-  "total_transitions": 100990260
+  "unique_contexts": 40202890,
+  "total_transitions": 75312653
 }

models/word_markov/ast_markov_ctx4_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:edd90f355e0db9ed093dd1b20ea832c072fe6c4eb8eda9897b82053102ed9e22
-size 1203044127
+oid sha256:9217b205ffd93647009467d670ec677b3039ae8a9d817ec195bc651d54e1f8c6
+size 1060990691

models/word_markov/ast_markov_ctx4_word_metadata.json

@@ -2,6 +2,6 @@
   "context_size": 4,
   "variant": "word",
   "language": "ast",
-  "unique_contexts": 66322442,
-  "total_transitions": 100852034
+  "unique_contexts": 57817277,
+  "total_transitions": 75176089
 }

models/word_ngram/ast_2gram_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4498d3f067003db47603a9cd549742083283922e1e3bcfac7e995862e92c13a0
-size 22263161
+oid sha256:6de399f50958b26f56a582db88426ccf2fe013228b455ccdf670d96c90d70911
+size 19677216

models/word_ngram/ast_2gram_word_metadata.json

@@ -2,6 +2,6 @@
   "n": 2,
   "variant": "word",
   "language": "ast",
-  "unique_ngrams": 1568799,
-  "total_ngrams": 101266716
+  "unique_ngrams": 1354323,
+  "total_ngrams": 75585781
 }

models/word_ngram/ast_3gram_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:77d87442f951bfee562611f31c08d2afebd188296c7dc1f3fc4691abc873c580
-size 61024149
+oid sha256:d09ec9dd887391dcc7d6cf6c2c1d8f84dece68e3a9cba2f5b80d23407481818d
+size 46625187

models/word_ngram/ast_3gram_word_metadata.json

@@ -2,6 +2,6 @@
   "n": 3,
   "variant": "word",
   "language": "ast",
-  "unique_ngrams": 3974147,
-  "total_ngrams": 101128488
+  "unique_ngrams": 2908394,
+  "total_ngrams": 75449217
 }

models/word_ngram/ast_4gram_word.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5a09cc0505f23b8e2addb0fd5ea158db1477037a7668eedab235820dfbcb3df5
-size 117588970
+oid sha256:a8515d5c220746ef576b0015d60dd4cc911b9af09f89293da61a12de1b2fbdb6
+size 80492927

models/word_ngram/ast_4gram_word_metadata.json

@@ -2,6 +2,6 @@
   "n": 4,
   "variant": "word",
   "language": "ast",
-  "unique_ngrams": 7247181,
-  "total_ngrams": 100990260
+  "unique_ngrams": 4707856,
+  "total_ngrams": 75312653
 }