mlsanityData to 7 Oct 2026

When a Name Is Not a Name: A Benchmark Dataset and Distilled Reasoning for Culturally Entangled Bangla Homographs in Low-Resource LLMs

arXiv 2607.17828 ↗First posted 20 Jul 2026cs.CLcited by 0

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

Many Bangla words are at once personal names and culturally loaded common nouns, "Maya" is both a girl's name and a word for affectionate compassion. Choosing the right reading demands cultural knowledge that is scarce in the pretraining data of modern language models. We introduce Culturally Entangled Homograph (CEH) disambiguation and build a Bangla benchmark of 1,516 expert-verified sentences (3,032 labelled occurrences) in which one word appears twice with two distinct readings, each labelled with a culturally grounded category and an explanation of the reasoning behind it. Across open- and closed-source models, we find a systematic dominant-meaning bias: models default to the common-noun sense and overlook the name. A Bangla-specific model fails under every prompting regime we test, showing that language-specific pretraining alone does not confer cultural grounding. We further show that contrastive chain-of-thought prompting can sharply reduce this bias without training, and that distilling cultural explanations teaches small (1-3B) models to reason toward the correct reading rather than memorise labels, cutting dominant-meaning bias from as high as 100% to under 5% and turning the failed Bangla-specific model into our strongest system. Dataset and code are available at https://github.com/ashuvo25/BanglaCEH.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 6 of its 25 references that are in the map. First posted on arXiv 20 Jul 2026.

Builds on

6 of 25 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025CulTrace: Tracing Internal Cultural Reasoning in Large Language Models12 Aug

Other references · 5

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  2. 2019BERTScore: Evaluating Text Generation with BERTcited by 5,462
  3. 2023Linguistic ambiguity analysis in ChatGPTcited by 14
  4. 2026Polite on the Surface, Broken in Practice: A Curated Dataset for Fixing Generation and Register Failures in Low-Resource Bangla Text Generationcited by 1
  5. 2026BenHalluEval: A Multi-Task Hallucination Evaluation Framework for Large Language Models on Bengalicited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025BengaliFig: A Low-Resource Challenge for Figurative and Culturally Grounded Reasoning in Bengali
  2. 2025Culturally-Grounded Chain-of-Thought (CG-CoT):Enhancing LLM Performance on Culturally-Specific Tasks in Low-Resource Languages
  3. 2025Can LLMs Solve My Grandma's Riddle? Evaluating Multilingual Large Language Models on Reasoning Traditional Bangla Tricky Riddles
  4. 2026XCR-Bench: Benchmarking Cross-Cultural Reasoning in LLMs via Culture-Specific Items and Hall's Triad
  5. 2025CRaFT: An Explanation-Based Framework for Evaluating Cultural Reasoning in Multilingual Language Models
  6. 2022On Reality and the Limits of Language Data: Aligning LLMs with Human Norms
  7. 2025From Word to World: Evaluate and Mitigate Culture Bias in LLMs via Word Association Test
  8. 2025Identity-Aware Large Language Models require Cultural Reasoning