arXiv subjects

Vivek Patel

Publications and source records attributed to Vivek Patel.

5 recordsLinked to original sources

arXiv2026

SectEval: Evaluating the Latent Sectarian Preferences of Large Language Models

As Large Language Models (LLMs) becomes a popular source for religious knowledge, it is important to know if it treats different groups fairly. This study is the first to measure how LLMs handle the differences between the two main sects of Islam: Sunni and Shia. We present a test called SectEval, available in both English and Hindi, consisting of 88 questions, to check the bias-ness of 15 top LLM models, both proprietary and open-weights. Our results show a major inconsistency based on language. In English, many powerful models DeepSeek-v3 and GPT-4o often favored Shia answers. However, when asked the exact same questions in Hindi, these models switched to favoring Sunni answers. This means a user could get completely different religious advice just by changing languages. We also looked at how models react to location. Advanced models Claude-3.5 changed their answers to match the user's country-giving Shia answers to a user from Iran and Sunni answers to a user from Saudi Arabia. In contrast, smaller models (especially in Hindi) ignored the user's location and stuck to a Sunni viewpoint. These findings show that AI is not neutral; its religious ``truth'' changes depending on the language you speak and the country you claim to be from. The data set is available at https://github.com/secteval/SectEval/

cs.CL↗

arXiv2025

IndicParam: Benchmark to evaluate LLMs on low-resource Indic Languages

While large language models excel on high-resource multilingual tasks, low- and extremely low-resource Indic languages remain severely under-evaluated. We present IndicParam, a human-curated benchmark of over 13,000 multiple-choice questions covering 11 such languages (Nepali, Gujarati, Marathi, Odia as low-resource; Dogri, Maithili, Rajasthani, Sanskrit, Bodo, Santali, Konkani as extremely low-resource) plus Sanskrit-English code-mixed set. We evaluated 20 LLMs, both proprietary and open-weights, which reveals that even the top-performing \texttt{Gemini-2.5} reaches 58\% average accuracy, followed by \texttt{GPT-5} (45) and \texttt{DeepSeek-3.2} (43.1). We additionally label each question as knowledge-oriented or purely linguistic to discriminate factual recall from grammatical proficiency. Further, we assess the ability of LLMs to handle diverse question formats-such as list-based matching, assertion-reason pairs, and sequence ordering-alongside conventional multiple-choice questions. \benchmark\ provides insights into limitations of cross-lingual transfer and establishes a challenging benchmark for Indic languages. The dataset is available at https://huggingface.co/datasets/bharatgenai/IndicParam. Scripts to run benchmark are present at https://github.com/ayushbits/IndicParam.

cs.CL↗

arXiv2025

Can Large Language Models Simulate Symbolic Execution Output Like KLEE?

Symbolic execution helps check programs by exploring different paths based on symbolic inputs. Tools like KLEE are commonly used because they can automatically detect bugs and create test cases. But one of KLEE's biggest issues is how slow it can get when programs have lots of branching paths-it often becomes too resource-heavy to run on large or complex code. In this project, we wanted to see if a large language model like GPT-4o could simulate the kinds of outputs that KLEE generates. The idea was to explore whether LLMs could one day replace parts of symbolic execution to save time and resources. One specific goal was to have GPT-4o identify the most constrained path in a program, this is the execution path with the most symbolic conditions. These paths are especially important because they often represent edge cases that are harder to test and more likely to contain deep bugs. However, figuring this out usually requires fully running KLEE, which can be expensive. So, we tested whether GPT-4o could predict the KLEE outputs and the most complex path using a dataset of 100 C programs. Our results showed about 20% accuracy in generating KLEE-like outputs and identifying the most constrained path. While not highly accurate, this early work helps show what current LLMs can and can't do when it comes to simulating symbolic execution.

cs.SE↗

arXiv2025

Snail Homing and Mating Search Algorithm for Weight Optimization of Stepped-Transmission Shaft

In this paper, the steeped-transmission shaft design problem is proposed for weight optimization. The bio-inspired search-based Snail Homing and Mating Search (SHMS) algorithm is utilized to solve the problem. It is inspired by the social behaviour of snails and their inherent nature of finding better homes, and mate. The proposed steeped-transmission shaft design problem is modelled considering the fatigue loading, combined bending, torsion loads, and the principle of Modified Goodman criteria. The forces diagram and the bending moment diagrams are obtained using the MDSOLIDS software. The forces and bending moment are then used to mathematical model the objective function and constraints. The SHMS algorithm has yielded the desired solution with reasonable computational cost. The constraints are handled using a static penalty function approach. The statistical results obtained using SHMS algorithm are further used for generating CAD model. The analysis is carried out in ANSYS Workbench. Further, the deflection obtained from SHMS algorithm and ANSYS Workbench are compared and results are discussed in details.

cs.NE↗

arXiv2025

ParamBench: A Graduate-Level Benchmark for Evaluating LLM Understanding on Indic Subjects

Large language models have been widely evaluated on tasks such as comprehension, summarization, code generation, etc. However, their performance on graduate-level, culturally grounded questions in the Indian context remains largely unexplored. Existing Indian benchmarks emphasise basic fact-orientated queries that offer limited assessment of a deeper disciplinary understanding tailored to the Indian setting. In this paper, we present ParamBench, consisting of more than 17K questions in the Hindi language, comprising questionnaires from 21 diverse subjects. These questions are primarily derived from a nationwide graduate-level entrance examination covering topics such as history, music, instruments, yoga, literature, philosophy, law, etc.~ specifically for the Indian context. Additionally, we assess the ability of LLMs to handle diverse question formats - such as list-based matching, assertion-reason pairs, and sequence ordering - alongside conventional multiple-choice questions. We evaluated the performance of more than 16 open source LLMs on this benchmark, observing that Gemma3-27B attains the highest overall accuracy of 56.4\%. Furthermore, subject-wise analysis indicates that even for the best-performing LLMs, performance remains weak on topics such as music, classical instruments, and law, underscoring persistent challenges in culturally grounded reasoning. The dataset and source code is present at https://github.com/ayushbits/ParamBench.

cs.CL↗