Benchmarking Bengali Dialectal Bias: A Multi-Stage Framework Integrating RAG-Based Translation and Human-Augmented RLAIF

• URL: http://arxiv.org/abs/2603.21359v1
• Date: Sun, 22 Mar 2026 18:44:57 GMT
• Title: Benchmarking Bengali Dialectal Bias: A Multi-Stage Framework Integrating RAG-Based Translation and Human-Augmented RLAIF
• Abstract summary: Large language models (LLMs) frequently exhibit performance biases against regional dialects of low-resource languages.<n>We propose a two-phase framework to evaluate dialectal bias in LLM question-answering across nine Bengali dialects.
• Score: 0.3227658251731014
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) frequently exhibit performance biases against regional dialects of low-resource languages. However, frameworks to quantify these disparities remain scarce. We propose a two-phase framework to evaluate dialectal bias in LLM question-answering across nine Bengali dialects. First, we translate and gold-label standard Bengali questions into dialectal variants adopting a retrieval-augmented generation (RAG) pipeline to prepare 4,000 question sets. Since traditional translation quality evaluation metrics fail on unstandardized dialects, we evaluate fidelity using an LLM-as-a-judge, which human correlation confirms outperforms legacy metrics. Second, we benchmark 19 LLMs across these gold-labeled sets, running 68,395 RLAIF evaluations validated through multi-judge agreement and human fallback. Our findings reveal severe performance drops linked to linguistic divergence. For instance, responses to the highly divergent Chittagong dialect score 5.44/10, compared to 7.68/10 for Tangail. Furthermore, increased model scale does not consistently mitigate this bias. We contribute a validated translation quality evaluation method, a rigorous benchmark dataset, and a Critical Bias Sensitivity (CBS) metric for safety-critical applications.

Related papers

• Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs [51.56484100374058]
  Large Language Models (LLMs) demonstrate remarkable multilingual fluency, but their internal knowledge representations remain disproportionately biased toward high-resource languages.<n>We investigate whether these biases can be mitigated at inference time, forcing an English-prompted model to answer as if it were queried in target languages.<n>Experiments on Gemma 3 12B Instruct reveal persona prompting to be the strongest overall intervention, balancing efficacy, safety, and out-of-domain generalization.
  arXiv  Detail & Related papers  (2026-07-21T16:15:05Z)
• Lower-Resource, Higher Scores: Language Bias in LLM Evaluators [27.959960538188753]
  We conduct experiments with semantically identical instruction-response pairs across 23 languages.<n>We find that multilingual evaluators assign significantly different scores to different evaluation languages.
  arXiv  Detail & Related papers  (2026-07-16T01:53:09Z)
• Retrieval-Augmented Linguistic Calibration [57.41519309308438]
  We model linguistic confidence as a distribution over plausible perceived probability values that a statement is correct.<n>We present Retrieval-Augmented Linguistic truth (RALC), a lightweight post-hoc pipeline that propagates confidence signals back into natural language.
  arXiv  Detail & Related papers  (2026-05-19T04:31:38Z)
• XQ-MEval: A Dataset with Cross-lingual Parallel Quality for Benchmarking Translation Metrics [64.77152900881724]
  We propose XQ-MEval, a semi-automatically built dataset covering nine translation directions.<n>We inject MQM-defined errors into gold translations automatically, filter them by native speakers for reliability, and merge errors to generate pseudo translations with controllable quality.<n>Using XQ-MEval, our experiments on nine representative metrics reveal the inconsistency between averaging and human judgment.
  arXiv  Detail & Related papers  (2026-04-16T12:27:10Z)
• Context-Aware Dialectal Arabic Machine Translation with Interactive Region and Register Selection [0.0]
  Current Machine Translation (MT) systems for Arabic often struggle to account for dialectal diversity.<n>We propose a context-aware framework for dialectal Arabic MT that explicitly models regional andlinguistic variation.
  arXiv  Detail & Related papers  (2026-04-07T20:54:19Z)
• Cross-Lingual Stability of LLM Judges Under Controlled Generation: Evidence from Finno-Ugric Languages [0.22009842278462158]
  Cross-lingual evaluation of large language models (LLMs) typically conflates two sources of variance: genuine model performance differences and measurement instability.<n>We investigate evaluation reliability by holding generation conditions constant while varying target language.<n>Our findings suggest that zero-shot judge transfer is unreliable for discourse-level assessment in morphologically rich languages.
  arXiv  Detail & Related papers  (2026-02-02T16:27:32Z)
• A Comparative Analysis of Retrieval-Augmented Generation Techniques for Bengali Standard-to-Dialect Machine Translation Using LLMs [0.3227658251731014]
  Translating from a standard language to its regional dialects is a significant NLP challenge due to scarce data and linguistic variation.<n>This paper proposes and compares two novel RAG pipelines for standard-to-dialectal Bengali translation.
  arXiv  Detail & Related papers  (2025-12-16T08:18:18Z)
• Steering LLMs toward Korean Local Speech: Iterative Refinement Framework for Faithful Dialect Translation [17.99472063920348]
  DIA-REFINE is a framework for goal-directed, inclusive dialect translation.<n>We introduce the dialect fidelity score (DFS) to quantify linguistic shift and the target dialect ratio (TDR) to measure the success of dialect translation.<n>Our work establishes a robust framework for goal-directed, inclusive dialect translation.
  arXiv  Detail & Related papers  (2025-11-10T03:52:24Z)
• PakBBQ: A Culturally Adapted Bias Benchmark for QA [3.4455728937232597]
  We introduce PakBBQ, a culturally and regionally adapted extension of the original Bias Benchmark for Question Answering dataset.<n> PakBBQ comprises over 214 templates, 17180 QA pairs across 8 categories in both English and Urdu, covering eight bias dimensions including age, disability, appearance, gender, socio-economic status, religious, regional affiliation, and language formality that are relevant in Pakistan.
  arXiv  Detail & Related papers  (2025-08-13T20:42:44Z)
• LLM-Based Evaluation of Low-Resource Machine Translation: A Reference-less Dialect Guided Approach with a Refined Sylheti-English Benchmark [1.3927943269211591]
  We propose a comprehensive framework that enhances Large Language Models (LLMs)-based machine translation evaluation.<n>We extend the ONUBAD dataset by incorporating Sylheti-English sentence pairs, corresponding machine translations, and Direct Assessment (DA) scores annotated by native speakers.<n>Our evaluation shows that the proposed pipeline consistently outperforms existing methods, achieving the highest gain of +0.1083 in Spearman correlation.
  arXiv  Detail & Related papers  (2025-05-18T07:24:13Z)
• Assessing Dialect Fairness and Robustness of Large Language Models in Reasoning Tasks [68.33068005789116]
  We introduce ReDial, a benchmark containing 1.2K+ parallel query pairs in Standardized English and AAVE.<n>We evaluate widely used models, including GPT, Claude, Llama, Mistral, and the Phi model families.<n>Our work establishes a systematic and objective framework for analyzing LLM bias in dialectal queries.
  arXiv  Detail & Related papers  (2024-10-14T18:44:23Z)
• CIF-Bench: A Chinese Instruction-Following Benchmark for Evaluating the Generalizability of Large Language Models [53.9835961434552]
  We introduce the Chinese Instruction-Following Benchmark (CIF-Bench) to evaluate the generalizability of large language models (LLMs) to the Chinese language. CIF-Bench comprises 150 tasks and 15,000 input-output pairs, developed by native speakers to test complex reasoning and Chinese cultural nuances. To mitigate data contamination, we release only half of the dataset publicly, with the remainder kept private, and introduce diversified instructions to minimize score variance.
  arXiv  Detail & Related papers  (2024-02-20T16:02:12Z)
• "Knowing When You Don't Know": A Multilingual Relevance Assessment Dataset for Robust Retrieval-Augmented Generation [90.09260023184932]
  Retrieval-Augmented Generation (RAG) grounds Large Language Model (LLM) output by leveraging external knowledge sources to reduce factual hallucinations. NoMIRACL is a human-annotated dataset for evaluating LLM robustness in RAG across 18 typologically diverse languages. We measure relevance assessment using: (i) hallucination rate, measuring model tendency to hallucinate, when the answer is not present in passages in the non-relevant subset, and (ii) error rate, measuring model inaccuracy to recognize relevant passages in the relevant subset.
  arXiv  Detail & Related papers  (2023-12-18T17:18:04Z)
• Benchmarking Large Language Models in Retrieval-Augmented Generation [53.504471079548]
  We systematically investigate the impact of Retrieval-Augmented Generation on large language models. We analyze the performance of different large language models in 4 fundamental abilities required for RAG. We establish Retrieval-Augmented Generation Benchmark (RGB), a new corpus for RAG evaluation in both English and Chinese.
  arXiv  Detail & Related papers  (2023-09-04T08:28:44Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.