Large models have advanced in processing long input sequences, but their ability to consistently use information across extended contexts remains a challenge. Recent studies highlight a positional bias where models prioritize information at the beginning or end of the input while neglecting the middle, resulting in a U-shaped performance curve but this was limited to English. Whether this bias is universal or shaped by language-specific factors remains unclear. In this work, we investigate positional bias in Bangla, a widely spoken but computationally underrepresented language. To support this, we introduce a novel Bangla benchmark dataset, bnContextQA, specifically designed for long-context comprehension. The dataset comprises of 350 long-context QA instances, each paired with 30 context paragraphs, allowing controlled evaluation of information retrieval at different positions. Using this dataset, we assess the performance of LLMs on Bangla across varying passage positions, providing insights into cross-linguistic positional effects. The bnContextQA dataset is publicly available at https://github.com/labiba02/bnContextQA.git to support future research on long-context understanding in Bangla and multilingual LLMs.

Large Language Models show impressive results in many different applications, most notably in the context of question-answering and open dialog situations. However, it is still an open question how to use those models for task-oriented dialogs such as booking or customer information systems, and such. In this work, we propose Dynamic Prompting, an architecture for task-oriented dialog, integrating the benefits of Large Language Models and showcasing the approach on the MultiWOZ 2.2 dataset. Our architecture leads to a high task success rate, provides sensible and specific answers, and is resistant to hallucinations. Further, we show that Dynamic Prompting is able to answer questions that were not anticipated by the dialog systems designer and that it can correct several types of errors and other characteristics of the system.