fix: scrub API keys from evaluation scripts for security

chore: remove hardcoded API keys and replace with environment variable loading

feat: add batched evaluation script for Gemini to bypass rate limits

chore: switch default Gemini model to gemini-1.5-flash for better free-tier limits

chore: switch default Gemini evaluation model to gemini-3.1-flash-lite

fix: add smart rate limit handling parsing wait_time from 429 errors for Gemini eval

feat: add eval_gemini.py to evaluate Gemini models via API

feat: rewrite eval_gonyai.py as full Kaggle-ready script matching colab_eval.py harness

fix: address ACL reviewer issues — non-monotonic scaling, OOD label, Gonyai/Gemma/Aya paragraphs, Section ref

feat: scatter 31 handcrafted items across 4 modules → 3,031 total items

chore: track jpg files with lfs

feat: backfill 37 handcrafted items to reach exactly 3,000 items and update paper

fix: resolve Overleaf PDF image rendering and suppress hbox warnings

fix: add amsmath and amssymb to fix LaTeX compilation errors

feat: add comprehensive statistical, validation, and diagnostic scripts

docs: finalize dataset validity, leakage claims, and length bias disclosure

fix: dataset integrity - pruned 67 malformed items

docs: replace tokenizer hypothesis with empirical fragmentation data

docs: major scientific upgrade for ACL Findings-level submission

docs: final promotional language sweep

docs: finalize structural evaluation elements for ACL submission

docs: add hyperparams, scaling laws, and refine tone

docs: fix critical data inconsistency and expand limitations

docs: finalize ACL reviewer feedback

docs: major revision of paper to meet ACL standards

fix: implement conditional probability scoring

fix: shuffle MCQ candidates, fix length bias, add difficulty tracking

feat: adapt eval script for Kaggle environment and 19GB disk limits

fix: enforce GPU 0 device_map and enable CPU offload for 8-bit models to bypass conservative memory dispatcher

fix: use BitsAndBytesConfig for newer transformers and clear disk cache to prevent IO Error

fix: total_mem -> total_memory (correct PyTorch API)

feat: v4.0 crash-safe eval suite — checkpoint/resume, aggressive VRAM cleanup, 12 mainstream models

docs: rewrite paper to professional 2-column format with expanded analysis; feat: add Google Colab evaluation script

docs: fix broken Devanagari shaping by reverting to Noto Serif Devanagari and update PDF

docs: add compiled PDF using FreeSerif Devanagari

docs: swap fonts to FreeSerif and Liberation Serif for better local compilation support

build: add latexmkrc to forcibly override Overleaf compiler to XeLaTeX

docs: re-add XeLaTeX magic comment to force Overleaf compiler

docs: revert back to pure Devanagari Konkani with XeLaTeX

docs: fix pdflatex fontspec crash by using romanized Konkani

docs: add GomParam research paper (XeLaTeX)

fix: removed all duplicate candidates across the benchmark using rigorous multi-pass review

fix: scrubbed all english translation hints from json files

feat: Artisan Wave 8 completion. 3030 hand-crafted items without templates

docs: add python usage snippet and fix item count typo in schema section

docs: update shields.io badges to reflect 2,164 items and 8,206 words

fix: remove custom python loader and define yaml configs for HF dataset viewer

docs: formalize README language for academic presentation

feat: Wave 6 complete - 2,164 items, ALL 20 modules have 100+ items

feat: push to 1,439 items (Wave 5), add 5 new modules