GomParam-v1: First dedicated Konkani language benchmark

.gitattributes

@@ -0,0 +1 @@
+*.jpg filter=lfs diff=lfs merge=lfs -text

GomParam-v1.py

@@ -0,0 +1,190 @@
+"""
+GomParam-v1: First Dedicated Konkani Language Benchmark
+Borkar, Omdeep (2026). GomParam-v1: A Benchmark for Evaluating Language
+Understanding in Konkani. https://huggingface.co/datasets/omdeep22/GomParam-v1
+
+Usage:
+    from datasets import load_dataset
+    ds = load_dataset("omdeep22/GomParam-v1", "cloze")
+    ds = load_dataset("omdeep22/GomParam-v1", "morphology")
+    ds = load_dataset("omdeep22/GomParam-v1", "para_qa")
+    ds = load_dataset("omdeep22/GomParam-v1", "jokes_sayings")
+    ds = load_dataset("omdeep22/GomParam-v1", "dialect")
+    ds = load_dataset("omdeep22/GomParam-v1", "perplexity")
+"""
+
+import json
+import datasets
+
+_CITATION = """\
+@misc{borkar2026gomparam,
+  title        = {GomParam-v1: A Benchmark for Evaluating Language Understanding in Konkani},
+  author       = {Borkar, Omdeep},
+  year         = {2026},
+  howpublished = {\\url{https://huggingface.co/datasets/omdeep22/GomParam-v1}},
+  note         = {First dedicated Konkani language benchmark covering morphology,
+                  cloze, comprehension, dialect robustness, and perplexity evaluation.}
+}
+"""
+
+_DESCRIPTION = """\
+GomParam-v1 is the first dedicated benchmark for evaluating large language models
+on Konkani — a severely low-resource Indo-Aryan language spoken primarily in Goa,
+India (ISO 639-3: kok). The benchmark is named after Gomantak, the ancient name
+for Goa, and mirrors the structure of IndicParam for Indic language evaluation.
+
+All tasks are designed to test language understanding rather than world knowledge,
+making GomParam-v1 a fair evaluation of any model's Konkani linguistic competence
+regardless of pretraining data size or multilingual exposure.
+
+Modules:
+  - cloze:         25 fill-in-the-blank items (case marking, agreement, postpositions)
+  - morphology:    20 verb conjugation items (tense, aspect, agreement, causatives)
+  - para_qa:       12 paragraph comprehension items (factual extraction, inference)
+  - jokes_sayings: 16 Konkani proverbs and jokes (pragmatic/cultural understanding)
+  - dialect:       15 Goan vs Mangalorean Konkani sentence pairs (robustness)
+  - perplexity:    30 held-out sentences (diverse registers, no overlap with training)
+
+Scoring: All MCQ tasks use 4-choice log-likelihood evaluation (argmax).
+         Perplexity tasks use mean bits-per-token on held-out sentences.
+Random baseline: 25.0% for all MCQ tasks.
+"""
+
+_HOMEPAGE = "https://huggingface.co/datasets/omdeep22/GomParam-v1"
+_LICENSE = "CC BY 4.0"
+
+_BASE_URL = "https://huggingface.co/datasets/omdeep22/GomParam-v1/resolve/main/data/"
+
+_DATA_FILES = {
+    "cloze":         "cloze.jsonl",
+    "morphology":    "morphology.jsonl",
+    "para_qa":       "para_qa.jsonl",
+    "jokes_sayings": "jokes_sayings.jsonl",
+    "dialect":       "dialect.jsonl",
+    "perplexity":    "perplexity.jsonl",
+}
+
+
+class GomParamConfig(datasets.BuilderConfig):
+    def __init__(self, name, **kwargs):
+        super().__init__(name=name, version=datasets.Version("1.0.0"), **kwargs)
+
+
+class GomParam(datasets.GeneratorBasedBuilder):
+    """GomParam-v1: First dedicated Konkani language benchmark."""
+
+    BUILDER_CONFIGS = [
+        GomParamConfig(name=subset, description=f"GomParam-v1 — {subset} module")
+        for subset in _DATA_FILES
+    ]
+
+    DEFAULT_CONFIG_NAME = "cloze"
+
+    def _info(self):
+        name = self.config.name
+
+        if name == "cloze":
+            features = datasets.Features({
+                "id":         datasets.Value("string"),
+                "sentence":   datasets.Value("string"),
+                "candidates": datasets.Sequence(datasets.Value("string")),
+                "correct":    datasets.Value("int32"),
+                "category":   datasets.Value("string"),
+                "note":       datasets.Value("string"),
+            })
+
+        elif name == "morphology":
+            features = datasets.Features({
+                "id":         datasets.Value("string"),
+                "context":    datasets.Value("string"),
+                "candidates": datasets.Sequence(datasets.Value("string")),
+                "correct":    datasets.Value("int32"),
+                "category":   datasets.Value("string"),
+                "note":       datasets.Value("string"),
+            })
+
+        elif name == "para_qa":
+            features = datasets.Features({
+                "id":         datasets.Value("string"),
+                "passage":    datasets.Value("string"),
+                "question":   datasets.Value("string"),
+                "candidates": datasets.Sequence(datasets.Value("string")),
+                "correct":    datasets.Value("int32"),
+                "category":   datasets.Value("string"),
+            })
+
+        elif name == "jokes_sayings":
+            features = datasets.Features({
+                "id":         datasets.Value("string"),
+                "type":       datasets.Value("string"),
+                "text":       datasets.Value("string"),
+                "question":   datasets.Value("string"),
+                "candidates": datasets.Sequence(datasets.Value("string")),
+                "correct":    datasets.Value("int32"),
+                "category":   datasets.Value("string"),
+            })
+
+        elif name == "dialect":
+            features = datasets.Features({
+                "id":               datasets.Value("string"),
+                "goan_dev":         datasets.Value("string"),
+                "mang_dev":         datasets.Value("string"),
+                "gloss":            datasets.Value("string"),
+                "lexical_diff":     datasets.Value("bool"),
+            })
+
+        elif name == "perplexity":
+            features = datasets.Features({
+                "id":       datasets.Value("string"),
+                "text":     datasets.Value("string"),
+                "register": datasets.Value("string"),
+                "domain":   datasets.Value("string"),
+            })
+
+        else:
+            raise ValueError(f"Unknown config: {name}")
+
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=features,
+            homepage=_HOMEPAGE,
+            license=_LICENSE,
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager):
+        url = _BASE_URL + _DATA_FILES[self.config.name]
+        filepath = dl_manager.download(url)
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                gen_kwargs={"filepath": filepath},
+            )
+        ]
+
+    def _generate_examples(self, filepath):
+        with open(filepath, encoding="utf-8") as f:
+            for key, line in enumerate(f):
+                line = line.strip()
+                if not line:
+                    continue
+                item = json.loads(line)
+
+                # Normalise optional fields so schema is always consistent
+                if self.config.name == "dialect":
+                    item.setdefault("lexical_diff", False)
+                    # Drop optional keys not in schema
+                    item = {k: v for k, v in item.items()
+                            if k in ["id", "goan_dev", "mang_dev", "gloss", "lexical_diff"]}
+
+                if self.config.name in ("cloze", "morphology"):
+                    item.setdefault("note", "")
+                    item.setdefault("category", "")
+
+                if self.config.name == "para_qa":
+                    item.setdefault("category", "")
+
+                if self.config.name == "jokes_sayings":
+                    item.setdefault("category", "")
+
+                yield key, item

README.md

@@ -0,0 +1,290 @@
+---
+language:
+  - kok
+license: cc-by-4.0
+task_categories:
+  - text-classification
+  - question-answering
+  - fill-mask
+task_ids:
+  - multiple-choice-qa
+  - natural-language-inference
+  - language-modeling
+pretty_name: GomParam-v1
+size_categories:
+  - n<1K
+tags:
+  - konkani
+  - low-resource
+  - indic-languages
+  - nlp-benchmark
+  - morphology
+  - language-understanding
+  - goa
+  - devanagari
+dataset_info:
+  - config_name: cloze
+    splits:
+      - name: test
+        num_examples: 25
+  - config_name: morphology
+    splits:
+      - name: test
+        num_examples: 20
+  - config_name: para_qa
+    splits:
+      - name: test
+        num_examples: 12
+  - config_name: jokes_sayings
+    splits:
+      - name: test
+        num_examples: 16
+  - config_name: dialect
+    splits:
+      - name: test
+        num_examples: 15
+  - config_name: perplexity
+    splits:
+      - name: test
+        num_examples: 30
+---
+
+# GomParam-v1 — First Dedicated Konkani Language Benchmark
+
+**GomParam** (named after *Gomantak*, the ancient Sanskrit name for Goa) is the first
+comprehensive benchmark designed specifically to evaluate large language models on
+**Konkani** (ISO 639-3: `kok`) — a severely low-resource Indo-Aryan language spoken
+by approximately 2.5 million speakers, primarily in Goa, India.
+
+> 📄 **Companion model:** [Gonyai-TEO2](https://huggingface.co/omdeep22/Gonyai-teo2) —
+> 251M parameter Konkani LLM pretrained from scratch.
+> 📦 **Companion corpus:** [Konkani-Books-Corpus-v2](https://huggingface.co/datasets/omdeep22/Konkani-Books-Corpus-v2) — 86M token Konkani dataset.
+
+---
+
+## Motivation
+
+Existing Indic language benchmarks (IndicParam, MILU, IndicGenBench) contain minimal
+or no Konkani coverage, and those that do test **world knowledge about Konkani culture**
+rather than **Konkani language ability**. GomParam-v1 fills this gap by testing:
+
+- Morphological correctness (verb conjugation, agreement)
+- Syntactic competence (case marking, postpositions, participles)
+- Reading comprehension in Konkani
+- Cultural and pragmatic understanding (proverbs, jokes)
+- Dialect robustness (Goan vs. Mangalorean Konkani)
+
+**No world knowledge is required.** Every question is answerable from language
+understanding alone, making GomParam-v1 a fair test for any model regardless of
+its encyclopedic pretraining.
+
+---
+
+## Dataset Structure
+
+### Modules
+
+| Module | Items | Task | Scoring |
+|---|---|---|---|
+| `cloze` | 25 | Fill-in-the-blank (4-choice) | Log-likelihood MCQ |
+| `morphology` | 20 | Verb conjugation (4-choice) | Log-likelihood MCQ |
+| `para_qa` | 12 | Paragraph comprehension (4-choice) | Log-likelihood MCQ |
+| `jokes_sayings` | 16 | Proverb/joke meaning (4-choice) | Log-likelihood MCQ |
+| `dialect` | 15 | Goan vs Mangalorean sentence pairs | Perplexity consistency |
+| `perplexity` | 30 | Held-out sentences | Bits-per-token |
+| **Total** | **118** | | |
+
+**Random baseline:** 25.0% for all MCQ tasks (4-choice).
+
+### Cloze Item Format
+```json
+{
+  "id": "cloze_001",
+  "sentence": "तो उद्यां मुंबयीक ___ वता.",
+  "candidates": ["विमानान", "विमाना", "विमानाक", "विमानानी"],
+  "correct": 0,
+  "category": "case_marking",
+  "note": "instrumental case — travel by plane"
+}
+```
+
+### Morphology Item Format
+```json
+{
+  "id": "morph_001",
+  "context": "हावें काल एक पुस्तक",
+  "candidates": ["वाचलें", "वाचलो", "वाचली", "वाचतां"],
+  "correct": 0,
+  "category": "ergative_past",
+  "note": "1sg ergative + neuter object past"
+}
+```
+
+### Para QA Item Format
+```json
+{
+  "id": "para_001",
+  "passage": "गोंय हें भारताच्या पश्चिम दर्यादेगेर...",
+  "question": "गोंय भारताक केन्ना मेळ्ळें?",
+  "candidates": ["१९४७ वर्सा", "१९६१ वर्सा", "१९५० वर्सा", "१९७१ वर्सा"],
+  "correct": 1,
+  "category": "factual_extraction"
+}
+```
+
+### Dialect Item Format
+```json
+{
+  "id": "dialect_004",
+  "goan_dev": "आमी उद्यां येतलो.",
+  "mang_dev": "आमी फाल्यां येतलो.",
+  "gloss": "We will come tomorrow.",
+  "lexical_diff": true
+}
+```
+
+---
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+# Load individual modules
+cloze    = load_dataset("omdeep22/GomParam-v1", "cloze",         split="test")
+morph    = load_dataset("omdeep22/GomParam-v1", "morphology",    split="test")
+para     = load_dataset("omdeep22/GomParam-v1", "para_qa",       split="test")
+jokes    = load_dataset("omdeep22/GomParam-v1", "jokes_sayings", split="test")
+dialect  = load_dataset("omdeep22/GomParam-v1", "dialect",       split="test")
+ppl_sents = load_dataset("omdeep22/GomParam-v1", "perplexity",  split="test")
+```
+
+### Evaluation (log-likelihood MCQ)
+
+```python
+import torch
+import numpy as np
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+model_id = "omdeep22/Gonyai-teo2"
+tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained(model_id, trust_remote_code=True)
+model.eval()
+
+@torch.no_grad()
+def score_completion(prompt, completion):
+    full = prompt + " " + completion
+    p_ids = tokenizer.encode(prompt, return_tensors="pt")
+    f_ids = tokenizer.encode(full,   return_tensors="pt")
+    if f_ids.shape[1] <= p_ids.shape[1]:
+        return float("-inf")
+    logits = model(f_ids).logits
+    opt_start  = p_ids.shape[1]
+    opt_logits = logits[0, opt_start - 1:-1, :]
+    opt_targets = f_ids[0, opt_start:]
+    lp = torch.nn.functional.log_softmax(opt_logits, dim=-1)
+    return lp[range(len(opt_targets)), opt_targets].mean().item()
+
+# Evaluate cloze
+correct = 0
+for item in cloze:
+    scores = [score_completion(item["sentence"].replace("___", ""), c)
+              for c in item["candidates"]]
+    if np.argmax(scores) == item["correct"]:
+        correct += 1
+print(f"Cloze accuracy: {correct/len(cloze)*100:.2f}%")
+```
+
+---
+
+## Benchmark Results (GomParam-v1)
+
+Results from the original paper evaluation. All models evaluated with 0-shot
+log-likelihood MCQ. Higher is better for all columns except PPL (lower is better).
+
+| Model | Params | Training | PPL↓ | Cloze | Morph | Para QA | Joke/Say | Dialect | **Composite** |
+|---|---|---|---|---|---|---|---|---|---|
+| Random Baseline | — | — | — | 25.0% | 25.0% | 25.0% | 25.0% | — | 25.0% |
+| Qwen2.5-0.5B | 0.5B | Multilingual | — | 40.0% | 41.7% | 83.3% | 12.5% | 79.0% | 53.8% |
+| Gemma-2-2B | 2B | Multilingual | — | 33.3% | 41.7% | 100% | 37.5% | 68.1% | 53.7% |
+| Sarvam-1 | 2B | Indic incl. Konkani | — | 20.0% | 25.0% | 100% | 12.5% | 75.2% | 40.9% |
+| **Gonyai-TEO2** | **251M** | **Konkani only** | **—** | **40.0%** | **75.0%** | **83.3%** | **37.5%** | **75.7%** | **🏆 64.2%** |
+
+> **Key finding:** Gonyai-TEO2 (251M parameters, Konkani-only pretraining) achieves the
+> highest composite score despite being 8× smaller than Sarvam-1 and Gemma-2-2B.
+> Morphology accuracy (75%) demonstrates that dedicated monolingual pretraining
+> confers strong grammatical competence that multilingual models cannot match at
+> equivalent scale. Multilingual models retain an advantage on Para QA tasks
+> where passage-level reading comprehension partially substitutes for language depth.
+
+---
+
+## Linguistic Coverage
+
+**Script:** Devanagari (primary Goan Konkani script)
+
+**Grammatical phenomena tested in Cloze & Morphology:**
+- Ergative-absolutive alignment (transitive past tense)
+- Gender agreement (masculine / feminine / neuter)
+- Number agreement (singular / plural)
+- Tense-aspect (present, past, future, imperfective, pluperfect)
+- Causative constructions (direct and indirect)
+- Case marking (nominative, accusative, instrumental, genitive, locative)
+- Postpositions and adverbial particles
+- Conjunctive and temporal participles
+- Relative clause pronoun resolution
+- Negation scope
+
+**Dialect pairs cover:**
+- Lexical variation (पाणी vs उदक, शाळा vs इस्कोल, पयसे vs दुडू)
+- Phonological variation (माका vs म्हाका, हावें vs हांवें)
+- Dialectal synonyms for temporal adverbs (उद्यां vs फाल्यां)
+
+---
+
+## Construction Methodology
+
+All benchmark items were hand-crafted by a native Goan Konkani speaker with
+reference to:
+
+- *A Grammar of Konkani* (Sardessai, 1986)
+- Goa Konkani Akademi linguistic reference materials
+- Native speaker intuition for naturalness verification
+
+Items were designed following these principles:
+1. **Language-only answerability** — no item requires world knowledge
+2. **Distractor plausibility** — wrong options are grammatically related forms
+3. **Register diversity** — colloquial, narrative, descriptive, prescriptive
+4. **Domain diversity** — family, nature, education, culture, emotion, agriculture
+
+---
+
+## Citation
+
+If you use GomParam-v1 in your research, please cite:
+
+```bibtex
+@misc{borkar2026gomparam,
+  title        = {GomParam-v1: A Benchmark for Evaluating Language Understanding in Konkani},
+  author       = {Borkar, Omdeep},
+  year         = {2026},
+  howpublished = {\url{https://huggingface.co/datasets/omdeep22/GomParam-v1}},
+  note         = {First dedicated Konkani language benchmark. Companion to Gonyai-TEO2.}
+}
+```
+
+---
+
+## Related Resources
+
+| Resource | Link |
+|---|---|
+| Gonyai-TEO2 (companion model) | [omdeep22/Gonyai-teo2](https://huggingface.co/omdeep22/Gonyai-teo2) |
+| Konkani-Books-Corpus-v2 | [omdeep22/Konkani-Books-Corpus-v2](https://huggingface.co/datasets/omdeep22/Konkani-Books-Corpus-v2) |
+| Benchmark code (Kaggle) | [GomParam evaluation notebook](https://github.com/Omdeepb69) |
+
+---
+
+## License
+
+[CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) — Free to use with attribution.

data/cloze.jsonl

@@ -0,0 +1,25 @@
+{"id": "cloze_001", "sentence": "तो उद्यां मुंबयीक ___ वता.", "candidates": ["विमानान", "विमाना", "विमानाक", "विमानानी"], "correct": 0, "category": "case_marking", "note": "instrumental case — travel by plane"}
+{"id": "cloze_002", "sentence": "ती ___ गाता.", "candidates": ["गोड", "गोडान", "गोडाक", "गोडें"], "correct": 0, "category": "adverb", "note": "adjective as adverb modifying singing"}
+{"id": "cloze_003", "sentence": "भुरग्यांनी ___ खेळचें.", "candidates": ["भायर", "भायरान", "भायरात", "भायराक"], "correct": 0, "category": "locative", "note": "playing outside — locative adverb"}
+{"id": "cloze_004", "sentence": "आमी ___ जेवण खातो.", "candidates": ["सांजेर", "सांजेरान", "सांजेराक", "सांजेरी"], "correct": 0, "category": "temporal", "note": "in the evening — temporal adverb"}
+{"id": "cloze_005", "sentence": "तिणें माका एक पुस्तक ___.", "candidates": ["दिलें", "दिलो", "दिली", "दितां"], "correct": 0, "category": "verb_agreement", "note": "neuter object past tense agreement"}
+{"id": "cloze_006", "sentence": "हो मनीस खूब ___ आसा.", "candidates": ["चांगलो", "चांगली", "चांगलें", "चांगले"], "correct": 0, "category": "adjective_agreement", "note": "masculine adjective agreement with मनीस"}
+{"id": "cloze_007", "sentence": "आमकां ___ भूक लागल्या.", "candidates": ["खूब", "खूबान", "खूबाक", "खूबें"], "correct": 0, "category": "adverb", "note": "degree adverb — very hungry"}
+{"id": "cloze_008", "sentence": "ती शाळेंत ___ वता.", "candidates": ["चालत", "चालतां", "चालतो", "चालताना"], "correct": 0, "category": "participle", "note": "present participial — goes walking"}
+{"id": "cloze_009", "sentence": "माझ्या ___ एक भाव आसा.", "candidates": ["बापायक", "बापायचो", "बापायन", "बापाय"], "correct": 1, "category": "genitive", "note": "genitive postposition — my father has a brother"}
+{"id": "cloze_010", "sentence": "देवळाचो ___ खूब उंच आसा.", "candidates": ["कळस", "कळसाक", "कळसान", "कळसाचो"], "correct": 0, "category": "nominative", "note": "nominative subject — the temple spire"}
+{"id": "cloze_011", "sentence": "तुवें हें काम ___ केलें.", "candidates": ["बरोबर", "बरोबरान", "बरोबराक", "बरोबरी"], "correct": 0, "category": "adverb", "note": "manner adverb — done correctly"}
+{"id": "cloze_012", "sentence": "पावसाळ्यांत नदी ___ वाहता.", "candidates": ["भरून", "भरुनी", "भरोन", "भरताना"], "correct": 0, "category": "participle", "note": "conjunctive participle — flows full"}
+{"id": "cloze_013", "sentence": "हांगासर ___ मनीस येतात.", "candidates": ["खूब", "खूबशे", "खूबशीं", "खूबशो"], "correct": 1, "category": "quantifier", "note": "quantifier agreeing with plural animate"}
+{"id": "cloze_014", "sentence": "ताणें माका ___ सांगलें.", "candidates": ["सत्य", "सत्यान", "सत्याक", "सत्याचें"], "correct": 0, "category": "direct_object", "note": "direct object — told the truth"}
+{"id": "cloze_015", "sentence": "भुरगो ___ पडलो.", "candidates": ["धांवतना", "धांवताना", "धांवत", "धांवतो"], "correct": 1, "category": "temporal_participle", "note": "while running — temporal participle"}
+{"id": "cloze_016", "sentence": "ताका काम करपाचें ___ नाशिल्लें.", "candidates": ["मन", "मनान", "मनाक", "मनाचें"], "correct": 0, "category": "nominative", "note": "nominative — he had no desire to work"}
+{"id": "cloze_017", "sentence": "ती आयज शाळेक ___ गेली नां.", "candidates": ["म्हणून", "म्हणान", "म्हणताना", "म्हणत"], "correct": 0, "category": "negation", "note": "reason clause with negation"}
+{"id": "cloze_018", "sentence": "जो मनीस खूब बरो आसा ___ सगळे वळखतात.", "candidates": ["ताका", "ताणें", "तो", "ताचो"], "correct": 0, "category": "relative_clause", "note": "relative pronoun resolution — accusative"}
+{"id": "cloze_019", "sentence": "घरांत ___ कोणूच नाशिल्लें.", "candidates": ["आयज", "आयजान", "आयजाक", "आयजाचें"], "correct": 0, "category": "temporal", "note": "temporal adverb in negative sentence"}
+{"id": "cloze_020", "sentence": "ताणें पुस्तक वाचून ___ ठेवलें.", "candidates": ["सकयल", "सकयलान", "सकयलाक", "सकयलाचें"], "correct": 0, "category": "locative", "note": "placed it down — directional adverb"}
+{"id": "cloze_021", "sentence": "हे दोन भाव एकामेकां ___ मेळटात.", "candidates": ["सारके", "सारकें", "सारको", "सारकेआत"], "correct": 0, "category": "adjective_agreement", "note": "resemblance — the two brothers look like each other"}
+{"id": "cloze_022", "sentence": "ताका ___ जेवण आवडटा.", "candidates": ["तिखट", "तिखटान", "तिखटाक", "तिखटाचें"], "correct": 0, "category": "adjective", "note": "he likes spicy food — predicative adjective"}
+{"id": "cloze_023", "sentence": "वर्सांतल्यान एकदां ___ घरा येता.", "candidates": ["तो", "ताणें", "ताका", "ताचो"], "correct": 0, "category": "nominative", "note": "nominative subject — he comes home once a year"}
+{"id": "cloze_024", "sentence": "ताणें माझो हात ___ धरलो.", "candidates": ["घट्ट", "घट्टान", "घट्टाक", "घट्टें"], "correct": 0, "category": "adverb", "note": "manner adverb — held tightly"}
+{"id": "cloze_025", "sentence": "रानांत ___ आवाज आयलो.", "candidates": ["कसलोशो", "कसलो", "कसलीशी", "कसलेंशें"], "correct": 0, "category": "indefinite", "note": "some kind of sound — indefinite masculine"}

data/dialect.jsonl

@@ -0,0 +1,15 @@
+{"id": "dialect_001", "goan_dev": "हावें जेवण खाल्लें.", "mang_dev": "हांवे जेवण खेल्लें.", "gloss": "I ate food.", "lexical_diff": false, "phonological_diff": true}
+{"id": "dialect_002", "goan_dev": "तो घरा गेलो.", "mang_dev": "तो घरा गेलो.", "gloss": "He went home.", "lexical_diff": false, "phonological_diff": false}
+{"id": "dialect_003", "goan_dev": "ती बरीच हुशार आसा.", "mang_dev": "ती बरीच हुशार आसा.", "gloss": "She is quite intelligent.", "lexical_diff": false, "phonological_diff": false}
+{"id": "dialect_004", "goan_dev": "आमी उद्यां येतलो.", "mang_dev": "आमी फाल्यां येतलो.", "gloss": "We will come tomorrow.", "lexical_diff": true, "note": "उद्यां (Goan) vs फाल्यां (Mangalorean) for 'tomorrow'"}
+{"id": "dialect_005", "goan_dev": "तुका कितें जाय?", "mang_dev": "तुका कितें जाय?", "gloss": "What do you want?", "lexical_diff": false, "phonological_diff": false}
+{"id": "dialect_006", "goan_dev": "भुरगो रडता.", "mang_dev": "भुरगो रडता.", "gloss": "The child is crying.", "lexical_diff": false, "phonological_diff": false}
+{"id": "dialect_007", "goan_dev": "माका भूक लागल्या.", "mang_dev": "म्हाका भूक लागल्या.", "gloss": "I am hungry.", "lexical_diff": false, "phonological_diff": true, "note": "माका (Goan) vs म्हाका (Mangalorean) for 'to me'"}
+{"id": "dialect_008", "goan_dev": "हें घर मोठें आसा.", "mang_dev": "हें घर व्हड आसा.", "gloss": "This house is big.", "lexical_diff": true, "note": "मोठें (Goan/Marathi influence) vs व्हड (Mangalorean) for 'big'"}
+{"id": "dialect_009", "goan_dev": "पाणी कितलें थंड आसा!", "mang_dev": "उदक कितलें थंड आसा!", "gloss": "How cold the water is!", "lexical_diff": true, "note": "पाणी (Goan) vs उदक (Mangalorean/classical) for 'water'"}
+{"id": "dialect_010", "goan_dev": "शाळा सकाळीं उगडटा.", "mang_dev": "इस्कोल सकाळीं उगडटा.", "gloss": "School opens in the morning.", "lexical_diff": true, "note": "शाळा (Goan/Marathi) vs इस्कोल (Mangalorean/Portuguese escola) for 'school'"}
+{"id": "dialect_011", "goan_dev": "हावें काम केलें.", "mang_dev": "हांवें काम केल्लें.", "gloss": "I did the work.", "lexical_diff": false, "phonological_diff": true, "note": "हावें vs हांवें — first person pronoun variation"}
+{"id": "dialect_012", "goan_dev": "ताचें नांव कितें?", "mang_dev": "ताचें नांव कितें?", "gloss": "What is his name?", "lexical_diff": false, "phonological_diff": false}
+{"id": "dialect_013", "goan_dev": "आमी जेवपाक बसले.", "mang_dev": "आमी जेवपाक बसले.", "gloss": "We sat down to eat.", "lexical_diff": false, "phonological_diff": false}
+{"id": "dialect_014", "goan_dev": "ताणें खूब पयसे मेळयले.", "mang_dev": "ताणें खूब दुडू मेळयले.", "gloss": "He earned a lot of money.", "lexical_diff": true, "note": "पयसे (Goan/Hindi) vs दुडू (Mangalorean/classical) for 'money'"}
+{"id": "dialect_015", "goan_dev": "तो बरो मनीस आसा.", "mang_dev": "तो बरो मनीस आसा.", "gloss": "He is a good person.", "lexical_diff": false, "phonological_diff": false}

data/jokes_sayings.jsonl

@@ -0,0 +1,16 @@
+{"id": "joke_001", "type": "saying", "text": "उदकाचो घोट घेतल्यार बरें.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["पाणी पिल्यार भूक मेळटा", "रागार आशिल्यार थंड रावन विचार करचो", "उदक पिवचें आरोग्याक बरें", "पाणी कमी पिल्यार आजार जाता"], "correct": 1, "category": "proverb_meaning"}
+{"id": "joke_002", "type": "saying", "text": "घरचो वैद्य गुणकारी.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["घरचो डॉक्टर सगळ्यांत चांगलो", "घरचे उपाय आनी सल्लो खूब उपयोगाचे", "घरांत औशध दवरचें", "वैद्यक शिक्षण घेवचें"], "correct": 1, "category": "proverb_meaning"}
+{"id": "joke_003", "type": "saying", "text": "हाताच्या फातराक लागीं नाका.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["दगड उचलूं नका", "आपल्याजवळ आशिल्ल्या गोष्टीची किंमत करची", "जड वस्तू हातान धरूं नका", "पाथर हातान मारूं नका"], "correct": 1, "category": "proverb_meaning"}
+{"id": "joke_004", "type": "joke", "text": "शिक्षकान विचारलें: 'सांग, पाऊस केन्ना पडटा?' भुरग्यान जाप दिली: 'सर, जेन्ना छत्री विसरतां तेन्ना!'", "question": "ह्या विनोदाचो मर्म कितलो?", "candidates": ["छत्री नेल्यार पाऊस पडना", "पाऊस येवपाची खबर नासता", "छत्री विसरल्यार नेमकी पाऊस पडटा हें सगळ्यांक अनुभवाचें", "शिक्षकाक पावसाचें ज्ञान नाशिल्लें"], "correct": 2, "category": "joke_punchline"}
+{"id": "joke_005", "type": "joke", "text": "दादलो: 'डॉक्टर, हावें जेवलें काय झोपतां.' डॉक्टर: 'मग जेवणा आदीं झोय!'", "question": "डॉक्टराच्या जापेंत विनोद कित्याक आसा?", "candidates": ["डॉक्टर चुकीची जाप दिता", "जेवणा उपरांत झोपप ही एक सवय", "जेवणा आदीं झोयल्यार जेवणा उपरांत झोप येवची नाशिल्ली — शब्दाचेर खेळ", "झोप येवप म्हणजे आजार"], "correct": 2, "category": "joke_punchline"}
+{"id": "joke_006", "type": "saying", "text": "आपलें तें आपणाकूच.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["आपल्या वस्तू दुसऱ्याक दिवं नका", "आपलें काम आपणूच करचें, दुसऱ्यांचेर हांव्यास धरूं नका", "आपलें घर आपणाक प्रिय", "स्वार्थी आसचें"], "correct": 1, "category": "proverb_meaning"}
+{"id": "joke_007", "type": "saying", "text": "नाचपाक नाकां मेळ्ळें म्हूण अंगण वांकडें.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["अंगण स्वच्छ दवरचें", "नाच शिकपाक जागो मुख्य", "काम जमना तेन्ना दुसऱ्याक दोश दिवप", "नाचपाचो सराव करचो"], "correct": 2, "category": "proverb_meaning"}
+{"id": "joke_008", "type": "joke", "text": "भुरगो: 'आवय, हावें परिक्षेंत शंबर मेळयले!' आवय: 'खरें? कसले विशयांत?' भुरगो: 'गणितांत पन्नास आनी विज्ञानांत पन्नास!'", "question": "ह्या विनोदाचो मर्म कितलो?", "candidates": ["भुरगो गणितांत हुशार आसा", "भुरग्यान एकठांय केल्ले गुण शंबर सांगले, पूण प्रत्यक्ष प्रत्येक विशयांत उणे", "आवयक गणित कळना", "शंबर गुण मेळोवप खूब कठीण"], "correct": 1, "category": "joke_punchline"}
+{"id": "joke_009", "type": "saying", "text": "मागता त्याका देवूळ नाका, मागनासताना त्याका देवूळ.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["देवळांत वचप गरजेचें", "जेन्ना गरज आसता तेन्ना मदत मेळना, जेन्ना गरज नासता तेन्ना मेळटा", "देव सगळ्यांक मदत करता", "मागपाची सवय वायट"], "correct": 1, "category": "proverb_meaning"}
+{"id": "joke_010", "type": "joke", "text": "शेजारिणीन विचारलें: 'तुमचो भुरगो कितलें शिकला?' आवयन जाप दिली: 'सगळें शिकला — झोपप, खावप आनी खेळप!'", "question": "ह्या विनोदांत आवय कितें सुचोवपाक मागता?", "candidates": ["भुरगो खूब हुशार आसा", "भुरगो शाळेंत कांयच शिकलो नां — फक्त मौजमस्ती केली", "शेजारीण खोशाल जाली", "भुरगो आजारी आसा"], "correct": 1, "category": "joke_punchline"}
+{"id": "joke_011", "type": "saying", "text": "एका हातान टाळी वाजना.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["टाळी वाजोवप कठीण", "कसलेंय काम जावपाक दोन्ही वटेन सहकार्य जाय", "एका हातान खूब काम जाता", "आवाज करपाक दोन हात जाय"], "correct": 1, "category": "proverb_meaning"}
+{"id": "joke_012", "type": "joke", "text": "बापूय: 'तुवें परिक्षेंत नापास जालें म्हणून हावें तुका सायकल दिवची नाशिल्ली.' भुरगो: 'बापूय, आनी तुमी परिक्षेंत कितले मेळयल्यालात तेन्ना आजाबान तुमकां कार दिली?'", "question": "भुरग्याची जाप चतुर कित्याक आसा?", "candidates": ["बापायक सायकल मेळ्ळी नाशिल्ली", "बापायन आपलें भुतकाळांत केल्लें ताच्याच तर्काचेर उलटयलें", "आजाब खूब श्रीमंत आसा", "भुरगो परिक्षेंत पास जालो"], "correct": 1, "category": "joke_punchline"}
+{"id": "joke_013", "type": "saying", "text": "बोलचालीत गोडास आनी काळजांत खोट.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["गोड उलोवप बरें", "उलोवपांत गोड पूण मनांत वायट विचार दवरपी मनशाबद्दल", "काळीज दुखता तेन्ना गोड खावचें", "मिठाय खावल्यार बरें दिसता"], "correct": 1, "category": "proverb_meaning"}
+{"id": "joke_014", "type": "saying", "text": "कामा नासताना धा वाटो.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["काम करपाक खूब वाटो आसात", "काम नाशिल्यार मनीस आळशी जाता आनी भोटें भटकता", "धा मनशांनी काम करचें", "वाटो सोदपाक जाय"], "correct": 1, "category": "proverb_meaning"}
+{"id": "joke_015", "type": "joke", "text": "डॉक्टरान सांगलें: 'तुमकां विश्रांती जाय.' रोग्यान विचारलें: 'कितलो वेळ?' डॉक्टर: 'जितलो वेळ तुमची बायल बोलना तितलो!'", "question": "डॉक्टराच्या जापेंत विनोद कित्याक आसा?", "candidates": ["बायल कधींच उलयना", "डॉक्टरान सांगपाक नजो आशिल्लें", "विश्रांती आनी बायलेचें उलोवप हांचो विनोदी संबंध जोडलो", "रोगी बरो जालो"], "correct": 2, "category": "joke_punchline"}
+{"id": "joke_016", "type": "saying", "text": "रानांत गेल्यार भुतां दिसतात.", "question": "ह्या म्हणीचो अर्थ कितलो?", "candidates": ["रानांत वचप धोकादायक", "भित्र्या मनशाक खंयूय भिरांत दिसता", "रानांत खरीच भुतां आसतात", "रातीं भायर वचूं नका"], "correct": 1, "category": "proverb_meaning"}

data/morphology.jsonl

@@ -0,0 +1,20 @@
+{"id": "morph_001", "context": "हावें काल एक पुस्तक", "candidates": ["वाचलें", "वाचलो", "वाचली", "वाचतां"], "correct": 0, "category": "ergative_past", "note": "1sg ergative + neuter object past"}
+{"id": "morph_002", "context": "ती भुरगी आयज शाळेक", "candidates": ["गेली", "गेलो", "गेलें", "वच्ची"], "correct": 0, "category": "intransitive_past", "note": "3sg feminine intransitive past"}
+{"id": "morph_003", "context": "आमी फाल्यां गोंयाक", "candidates": ["वतात", "वतों", "वतो", "वतां"], "correct": 0, "category": "future_habitual", "note": "1pl future/habitual go"}
+{"id": "morph_004", "context": "तो आतां जेवण", "candidates": ["खाता", "खाती", "खातें", "खातो"], "correct": 0, "category": "present", "note": "3sg masculine present eat"}
+{"id": "morph_005", "context": "तिणें एक गाणें", "candidates": ["म्हटलें", "म्हटलो", "म्हटली", "म्हणटा"], "correct": 0, "category": "ergative_past", "note": "3sg feminine ergative + neuter object past"}
+{"id": "morph_006", "context": "भुरग्यांनी खेळ", "candidates": ["खेळलो", "खेळलीं", "खेळलें", "खेळतात"], "correct": 1, "category": "ergative_past_plural", "note": "3pl ergative past — neuter/mixed agreement"}
+{"id": "morph_007", "context": "हावें तुका एक गजाल", "candidates": ["सांगचें", "सांगतलों", "सांगता", "सांगूं"], "correct": 1, "category": "future", "note": "1sg future — I will tell you something"}
+{"id": "morph_008", "context": "ताणें माझें पुस्तक", "candidates": ["घेतलें", "घेतलो", "घेतली", "घेता"], "correct": 0, "category": "ergative_past", "note": "3sg masculine ergative + neuter object past"}
+{"id": "morph_009", "context": "तीं भुरगीं बागेंत", "candidates": ["खेळतालीं", "खेळताली", "खेळतालो", "खेळताला"], "correct": 0, "category": "imperfective_past", "note": "3pl neuter imperfective past"}
+{"id": "morph_010", "context": "तुमी काल कित्याक", "candidates": ["आयलात", "आयलो", "आयली", "आयलें"], "correct": 0, "category": "intransitive_past_plural", "note": "2pl intransitive past come"}
+{"id": "morph_011", "context": "माझी आवय स्वयंपाकघरांत जेवण", "candidates": ["रांदता", "रांदती", "रांदतो", "रांदतें"], "correct": 0, "category": "present", "note": "3sg feminine present cook"}
+{"id": "morph_012", "context": "हावें उद्यां एक खत", "candidates": ["बरयतलों", "बरयतलो", "बरयतली", "बरयतलें"], "correct": 0, "category": "future", "note": "1sg future write — ergative future"}
+{"id": "morph_013", "context": "ताणें भुरग्याक आंब्याचें झाड", "candidates": ["दाखयलें", "दाखयलो", "दाखयली", "दाखयता"], "correct": 0, "category": "causative_past", "note": "causative — showed the child the mango tree, neuter object"}
+{"id": "morph_014", "context": "आवयन भुरग्याक भात", "candidates": ["जेवयलो", "जेवयलें", "जेवयली", "जेवयता"], "correct": 0, "category": "causative_past", "note": "causative — mother fed the child rice, masculine object"}
+{"id": "morph_015", "context": "ताका बरें ___ उपरांत तो भायर सरलो।", "candidates": ["जालें", "जालो", "जाली", "जातां"], "correct": 0, "category": "impersonal", "note": "impersonal — after he felt better (neuter)"}
+{"id": "morph_016", "context": "पावसान झाडां", "candidates": ["पडयलीं", "पडयलो", "पडयलें", "पडयली"], "correct": 0, "category": "causative_past_plural", "note": "rain caused trees to fall — neuter plural object"}
+{"id": "morph_017", "context": "तुवें हें काम आदीं", "candidates": ["केल्लेंआसां", "केल्लोआसां", "केल्लीआसां", "करतां"], "correct": 0, "category": "perfect_pluperfect", "note": "pluperfect — you had done this work before"}
+{"id": "morph_018", "context": "ती दर सकाळीं फुलां", "candidates": ["काडटा", "काडतो", "काडतीं", "काडटात"], "correct": 0, "category": "habitual_present", "note": "3sg feminine habitual — picks flowers every morning"}
+{"id": "morph_019", "context": "ताणें घर बांदपाक मनीस", "candidates": ["लायले", "लायलें", "लायलो", "लायली"], "correct": 0, "category": "causative_plural_object", "note": "he got workers to build the house — plural animate object"}
+{"id": "morph_020", "context": "भुरग्यान दुदू", "candidates": ["पियेलें", "पियेलो", "पियेली", "पितां"], "correct": 0, "category": "ergative_past_liquid", "note": "child drank milk — neuter liquid object ergative past"}

data/para_qa.jsonl

@@ -0,0 +1,12 @@
+{"id": "para_001", "passage": "गोंय हें भारताच्या पश्चिम दर्यादेगेर आशिल्लें एक सुंदर राज्य आसा. हांगा खूब सुंदर दर्यावेळो आसा. गोंयचे लोक कोंकणी भाशेंत उलयतात. हांगा पोर्तुगेजांनी खूब वर्सां राज्य केलें. गोंय १९६१ वर्सा भारताक मेळ्ळें.", "question": "गोंय भारताक केन्ना मेळ्ळें?", "candidates": ["१९४७ वर्सा", "१९६१ वर्सा", "१९५० वर्सा", "१९७१ वर्सा"], "correct": 1, "category": "factual_extraction"}
+{"id": "para_002", "passage": "कोंकणी भास ही गोंयची राजभास आसा. ही भास देवनागरी लिपयेंत बरयतात. कोंकणी भाशेंत खूब गोड कविता आनी गाणीं आसात. ह्या भाशेक भारतीय संविधानाच्या आठव्या अनुसुचींत स्थान मेळ्ळां.", "question": "कोंकणी भास कसली लिपी वापरता?", "candidates": ["रोमी लिपी", "देवनागरी लिपी", "कन्नड लिपी", "मल्याळम लिपी"], "correct": 1, "category": "factual_extraction"}
+{"id": "para_003", "passage": "दर्यावेळच्या वाऱ्यान नारळाचीं पानां हालतालीं. रामूने आपल्या बापायसवें मासो धरपाक गेलो. तांणी सकाळीं पांच वाजतां निघाले. दनपारां ताणी खूब मासो धरलो आनी घरा परतले.", "question": "रामू कितल्या वाजतां निघालो?", "candidates": ["सहा वाजतां", "सात वाजतां", "पांच वाजतां", "चार वाजतां"], "correct": 2, "category": "factual_extraction"}
+{"id": "para_004", "passage": "सरिता एक हुशार भुरगी आसा. ती दर वर्सा परिक्षेंत पयलें येता. तिची आवय शिक्षिका आसा आनी बापूय डॉक्टर आसा. सरिताक फुडें इंजिनियर जावचें आसा.", "question": "सरिताच्या बापायचो धंदो कितलो?", "candidates": ["शिक्षक", "इंजिनियर", "डॉक्टर", "वकील"], "correct": 2, "category": "factual_extraction"}
+{"id": "para_005", "passage": "गोंयांत मांडो हो एक लोकप्रिय नाच आसा. हो नाच पोर्तुगेज काळांत सुरू जालो. ह्या नाचांत बायलो आनी दादले एकत्र नाचतात. मांडो नाचाक गोंयच्या संस्कृतीचें प्रतीक मानतात.", "question": "मांडो नाच केन्ना सुरू जालो?", "candidates": ["मराठा काळांत", "ब्रिटीश काळांत", "पोर्तुगेज काळांत", "मुघल काळांत"], "correct": 2, "category": "factual_extraction"}
+{"id": "para_006", "passage": "काजू हें गोंयचें एक मुख्य पीक आसा. काजूच्या फळापसून फेणी हें पेय तयार करतात. काजूचो हंगाम फेब्रुवारी ते एप्रिल मेनेजांत आसता. गोंयांत काजूचें उत्पादन खूब जाता.", "question": "काजूचो हंगाम कितल्या म्हयन्यांत आसता?", "candidates": ["जून ते ऑगस्ट", "फेब्रुवारी ते एप्रिल", "ऑक्टोबर ते डिसेंबर", "नोव्हेंबर ते जानेवारी"], "correct": 1, "category": "factual_extraction"}
+{"id": "para_007", "passage": "विमलाबायेक स्वयंपाक करपाची खूब आवड आसली. ती दर आयतारा घरच्यांखातीर खास जेवण बनयताली. ताचे घोव आनी भुरगे ताचो स्वयंपाक खूब मेळयतालें. पूण एक दीस ती आजारी पडल�� आनी उठपाक जमलें नाशिल्लें.", "question": "विमलाबाय केन्ना खास जेवण बनयताली?", "candidates": ["शेनवारा", "आयतारा", "सोमारा", "बुधवारा"], "correct": 1, "category": "factual_extraction"}
+{"id": "para_008", "passage": "मोहनान आपल्या शेतांत ह्या वर्सा नवें बी पेरलें. पाऊस बरो पडलो आनी पीक खूब बरें आयलें. ताणें बाजारांत पीक विकलें आनी बरे पयसे मेळयले. त्या पयशांनी ताणें आपल्या भुरग्यांखातीर नव्यो चपली घेतल्यो.", "question": "मोहनान पयशांनी कितें घेतलें?", "candidates": ["नवें बी", "खत", "नव्यो चपली", "गाय"], "correct": 2, "category": "inference"}
+{"id": "para_009", "passage": "रेखान आपल्या पोरसांत खूब वेगवेगळीं फुलां लायल्यांत. गुलाब, चाफो आनी मोगरो तांच्या आवडीचीं फुलां आसात. सकाळीं ती फुलां काडून देवाक वाहता. सांजेर ती पोरसांत बसून पुस्तक वाचता.", "question": "रेखा सकाळीं फुलां काडून कित्याक वापरता?", "candidates": ["घर सजोवपाक", "विकपाक", "देवाक वाहपाक", "भेटयेखातीर"], "correct": 2, "category": "inference"}
+{"id": "para_010", "passage": "ह्या गांवांत एक पुर्विल्लें देवूळ आसा. हें देवूळ पांचशें वर्सां पुर्विल्लें आसा असो लोक सांगतात. दर वर्सा हांगा व्हडली जात्रा भरता. जात्रेक गोंयाच्या वेगवेगळ्या भागांतल्यान लोक येतात.", "question": "देवूळ किते वर्सां पुर्विल्लें आसा?", "candidates": ["शंबर वर्सां", "दोनशें वर्सां", "पांचशें वर्सां", "हजार वर्सां"], "correct": 2, "category": "factual_extraction"}
+{"id": "para_011", "passage": "सागरान धा वर्सां कठोर मेहनत करून डॉक्टर जालो. ताच्या घरची परिस्थिती बरी नाशिल्ली. पूण ताची आवय ताका शिकोवपाखातीर खूब कश्ट केले. आयज सागर एका व्हड दवाखान्यांत काम करता आनी आपल्या आवयची खूब काळजी घेता.", "question": "सागराची शिकपाक मदत कोणें केली?", "candidates": ["बापायन", "आवयन", "मास्तरान", "सरकारान"], "correct": 1, "category": "inference"}
+{"id": "para_012", "passage": "कोंकणी भाशेंत दोन मुख्य लिप्यो आसात — देवनागरी आनी रोमी. गोंयांत देवनागरी लिपी सरकारी कामांखातीर वापरतात. मंगळुरांत रोमी लिपी चडशी वापरतात. दोनूय लिप्यो एकाच भाशेचें वेगवेगळें रूप दाखयतात.", "question": "गोंयांत सरकारी कामांखातीर कसली लिपी वापरतात?", "candidates": ["रोमी लिपी", "कन्नड लिपी", "देवनागरी लिपी", "मल्याळम लिपी"], "correct": 2, "category": "factual_extraction"}

data/perplexity.jsonl

@@ -0,0 +1,30 @@
+{"id": "ppl_001", "text": "आमी आयज बजारांत भाजी विकत घेतली.", "register": "colloquial", "domain": "daily_life"}
+{"id": "ppl_002", "text": "तो सकाळीं लवकर उठता आनी देवळाक वता.", "register": "colloquial", "domain": "daily_life"}
+{"id": "ppl_003", "text": "हांगा पाऊस खूब पडता, पूण आमकां बरें दिसता.", "register": "colloquial", "domain": "weather"}
+{"id": "ppl_004", "text": "ती भुरगी शाळेंत खूब हुशार आसा.", "register": "colloquial", "domain": "education"}
+{"id": "ppl_005", "text": "माझी आवय स्वयंपाकघरांत जेवण रांदता.", "register": "colloquial", "domain": "family"}
+{"id": "ppl_006", "text": "गोंयांत नारळाचीं झाडां खूब आसात.", "register": "descriptive", "domain": "nature"}
+{"id": "ppl_007", "text": "तुमी कसले काम करतात?", "register": "conversational", "domain": "daily_life"}
+{"id": "ppl_008", "text": "हे गांवांत लोक एकामेकाक वळखतात.", "register": "descriptive", "domain": "community"}
+{"id": "ppl_009", "text": "दर्यावेळी वारो खूब सुखाचो आसता.", "register": "descriptive", "domain": "nature"}
+{"id": "ppl_010", "text": "भुरग्यांनी खेळटना खूब आनंद मेळटा.", "register": "descriptive", "domain": "childhood"}
+{"id": "ppl_011", "text": "आमच्या घरांत पाच जण आसात.", "register": "colloquial", "domain": "family"}
+{"id": "ppl_012", "text": "तिणें एक सुंदर गाणें म्हटलें.", "register": "narrative", "domain": "culture"}
+{"id": "ppl_013", "text": "शेतकार सकाळीं उठून शेतार काम करता.", "register": "descriptive", "domain": "agriculture"}
+{"id": "ppl_014", "text": "पोरसांत फुलां फुल्लिल्लीं आसात.", "register": "descriptive", "domain": "nature"}
+{"id": "ppl_015", "text": "तो मनीस खूब दयाळू आसा.", "register": "colloquial", "domain": "character"}
+{"id": "ppl_016", "text": "आमकां उद्यां परतून येवचें आसा.", "register": "colloquial", "domain": "daily_life"}
+{"id": "ppl_017", "text": "ह्या वाटेन गेल्यार गांव मेळटा.", "register": "instructional", "domain": "navigation"}
+{"id": "ppl_018", "text": "तिची भास खूब गोड आसा.", "register": "colloquial", "domain": "language"}
+{"id": "ppl_019", "text": "लहान भुरग्यांनी आपल्या आयबापायकां मान दिवचो.", "register": "prescriptive", "domain": "values"}
+{"id": "ppl_020", "text": "सांजेर देवळांत घंटो वाजता.", "register": "descriptive", "domain": "culture"}
+{"id": "ppl_021", "text": "ताणें आपल्या बापायक बरें जेवण दिलें.", "register": "narrative", "domain": "family"}
+{"id": "ppl_022", "text": "पावसाळ्यांत नद्यो भरून वाहतात.", "register": "descriptive", "domain": "nature"}
+{"id": "ppl_023", "text": "गांवांतले लोक जात्रेक एकठांय येतात.", "register": "descriptive", "domain": "culture"}
+{"id": "ppl_024", "text": "ताका नव्या शाळेंत जावपाची भिरांत वाटताली.", "register": "narrative", "domain": "emotion"}
+{"id": "ppl_025", "text": "ती बायल दर दिसा पांच किलोमीटर चालता.", "register": "descriptive", "domain": "health"}
+{"id": "ppl_026", "text": "भुरग्यांनी परिक्षेखातीर खूब अभ्यास केलो.", "register": "narrative", "domain": "education"}
+{"id": "ppl_027", "text": "ह्या झाडाक आंबे खूब आसात ह्या वर्सा.", "register": "colloquial", "domain": "nature"}
+{"id": "ppl_028", "text": "वैजयंतीन आपल्या भावाखातीर जेवण बनयलें.", "register": "narrative", "domain": "family"}
+{"id": "ppl_029", "text": "रातीं आकाशांत चोंदवो बरोच दिसतालो.", "register": "descriptive", "domain": "nature"}
+{"id": "ppl_030", "text": "कोंकणी भास शिकपाक तांका खूब आवड आसा.", "register": "descriptive", "domain": "language"}