@inproceedings{shuvo-etal-2026-polite,
title = "Polite on the Surface, Broken in Practice: A Curated Dataset for Fixing Generation and Register Failures in Low-Resource {B}angla Text Generation",
author = "Shuvo, Md. Asaduzzaman and
Hasan, Mahedi and
Parvez, Md. Tashin and
Noman, Azizul Haque and
Ovi, Md. Shafayet Hossain",
editor = "Wang, Yimu and
Choi, Yee Man and
Wu, Di and
Ouyang, Siqi and
Azadani, Mozhgan Nasr and
Fung, Yi R. (May)",
booktitle = "Proceedings of the 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently ({G}round{LM} 2026)",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.groundlm-1.56/",
pages = "751--766",
ISBN = "979-8-89176-478-1",
abstract = "Recent advances in Multilingual Large Language Models (MLLMs) have improved cross-lingual conversational capabilities, yet generating reliable, institutionally usable text in low-resource languages like Bangla remains a critical bottleneck, particularly for compact, low-resource LLMs. Beyond surface fluency, state-of-the-art and small open-weight models frequently exhibit deeper generation failures in Bangla: incoherent or off-topic output, structural misformatting, and inconsistent honorific registers, that render text unusable in formal contexts. To address this, we introduce BLADE (BangLa Application and DialoguE generation), a culturally aligned instruction-tuning dataset and benchmarking framework comprising 4,196 curated interaction pairs. We use this resource to fine-tune and evaluate compact open-weight architectures, including DeepSeek-8B, LLaMA-3.2-3B, Qwen2, and TigerLLM-1B, via LoRA adapters under 4-bit NormalFloat (NF4) quantization on limited hardware. Our evaluations show that targeted fine-tuning on BLADE yields substantial improvements in generation reliability, structural fidelity, and register consistency even for the smallest models, establishing a rigorous benchmark for closing generation and pragmatic gaps in low-resource multilingual text generation. Code and dataset: https://github.com/ashuvo25/Bangla{\_}Application{\_}LLM"
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="shuvo-etal-2026-polite">
<titleInfo>
<title>Polite on the Surface, Broken in Practice: A Curated Dataset for Fixing Generation and Register Failures in Low-Resource Bangla Text Generation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Md.</namePart>
<namePart type="given">Asaduzzaman</namePart>
<namePart type="family">Shuvo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mahedi</namePart>
<namePart type="family">Hasan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Md.</namePart>
<namePart type="given">Tashin</namePart>
<namePart type="family">Parvez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Azizul</namePart>
<namePart type="given">Haque</namePart>
<namePart type="family">Noman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Md.</namePart>
<namePart type="given">Shafayet</namePart>
<namePart type="given">Hossain</namePart>
<namePart type="family">Ovi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently (GroundLM 2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yimu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yee</namePart>
<namePart type="given">Man</namePart>
<namePart type="family">Choi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Di</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Siqi</namePart>
<namePart type="family">Ouyang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mozhgan</namePart>
<namePart type="given">Nasr</namePart>
<namePart type="family">Azadani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="given">R</namePart>
<namePart type="given">(May)</namePart>
<namePart type="family">Fung</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-478-1</identifier>
</relatedItem>
<abstract>Recent advances in Multilingual Large Language Models (MLLMs) have improved cross-lingual conversational capabilities, yet generating reliable, institutionally usable text in low-resource languages like Bangla remains a critical bottleneck, particularly for compact, low-resource LLMs. Beyond surface fluency, state-of-the-art and small open-weight models frequently exhibit deeper generation failures in Bangla: incoherent or off-topic output, structural misformatting, and inconsistent honorific registers, that render text unusable in formal contexts. To address this, we introduce BLADE (BangLa Application and DialoguE generation), a culturally aligned instruction-tuning dataset and benchmarking framework comprising 4,196 curated interaction pairs. We use this resource to fine-tune and evaluate compact open-weight architectures, including DeepSeek-8B, LLaMA-3.2-3B, Qwen2, and TigerLLM-1B, via LoRA adapters under 4-bit NormalFloat (NF4) quantization on limited hardware. Our evaluations show that targeted fine-tuning on BLADE yields substantial improvements in generation reliability, structural fidelity, and register consistency even for the smallest models, establishing a rigorous benchmark for closing generation and pragmatic gaps in low-resource multilingual text generation. Code and dataset: https://github.com/ashuvo25/Bangla_Application_LLM</abstract>
<identifier type="citekey">shuvo-etal-2026-polite</identifier>
<location>
<url>https://aclanthology.org/2026.groundlm-1.56/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>751</start>
<end>766</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Polite on the Surface, Broken in Practice: A Curated Dataset for Fixing Generation and Register Failures in Low-Resource Bangla Text Generation
%A Shuvo, Md. Asaduzzaman
%A Hasan, Mahedi
%A Parvez, Md. Tashin
%A Noman, Azizul Haque
%A Ovi, Md. Shafayet Hossain
%Y Wang, Yimu
%Y Choi, Yee Man
%Y Wu, Di
%Y Ouyang, Siqi
%Y Azadani, Mozhgan Nasr
%Y Fung, Yi R. (May)
%S Proceedings of the 1st Workshop on Grounding Language Models: Learning Faithfully and Efficiently (GroundLM 2026)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-478-1
%F shuvo-etal-2026-polite
%X Recent advances in Multilingual Large Language Models (MLLMs) have improved cross-lingual conversational capabilities, yet generating reliable, institutionally usable text in low-resource languages like Bangla remains a critical bottleneck, particularly for compact, low-resource LLMs. Beyond surface fluency, state-of-the-art and small open-weight models frequently exhibit deeper generation failures in Bangla: incoherent or off-topic output, structural misformatting, and inconsistent honorific registers, that render text unusable in formal contexts. To address this, we introduce BLADE (BangLa Application and DialoguE generation), a culturally aligned instruction-tuning dataset and benchmarking framework comprising 4,196 curated interaction pairs. We use this resource to fine-tune and evaluate compact open-weight architectures, including DeepSeek-8B, LLaMA-3.2-3B, Qwen2, and TigerLLM-1B, via LoRA adapters under 4-bit NormalFloat (NF4) quantization on limited hardware. Our evaluations show that targeted fine-tuning on BLADE yields substantial improvements in generation reliability, structural fidelity, and register consistency even for the smallest models, establishing a rigorous benchmark for closing generation and pragmatic gaps in low-resource multilingual text generation. Code and dataset: https://github.com/ashuvo25/Bangla_Application_LLM
%U https://aclanthology.org/2026.groundlm-1.56/
%P 751-766
Markdown (Informal)
[Polite on the Surface, Broken in Practice: A Curated Dataset for Fixing Generation and Register Failures in Low-Resource Bangla Text Generation](https://aclanthology.org/2026.groundlm-1.56/) (Shuvo et al., GroundLM 2026)
ACL