Searcharxiv⌕ Search

arXiv subjects

Hari Sitaula

Publications and source records attributed to Hari Sitaula.

3 recordsLinked to original sources

Adapting Small Language Models to Romanized Nepali: Tokenizer Fertility, Reference Likelihood, and Generation Metrics

Romanized Nepali, Nepali written in Latin script, is common in digital communication but has no standard spelling and few resources, and it is unclear how reliably small open language models can be adapted to it. We adapt three 2-3B base models from different families, Llama-3.2-3B, Qwen2.5-3B, and Gemma-2-2B, with QLoRA on 9,000 instruction examples produced by deterministic transliteration of Nepali instruction data. Each model is trained with one shared recipe and three seeds and evaluated on 1,000 held-out examples with reference-overlap metrics, byte-normalized reference likelihood (bits per byte, BPB), and an output audit, against zero-shot baselines under two prompts. Fine-tuning raises chrF++ from 10-14 to 29.0-30.6 and lowers BPB from 2.23-2.41 to 0.78-0.84; most zero-shot outputs (54-92%) are English-like. Likelihood and overlap disagree on how far apart the models are. Llama-3.2-3B has the lowest BPB in every seed, 0.042 below Qwen2.5-3B (95% CI 0.039-0.046), yet the two cannot be separated on chrF++ (30.51 and 30.61) or BLEU, even with paired item-level tests. Gemma-2-2B, whose tokenizer splits Romanized Nepali words least (2.72 tokens per word), scores worst on every measure. Which model gains most over its zero-shot baseline depends on the zero-shot prompt. In a supplementary comparison with an identical tokenizer, Llama-3.1-8B lowers BPB by 0.053 (0.050-0.056) relative to Llama-3.2-3B, while chrF++ and BLEU rise by only 0.68 and 0.58 points. Lower fertility and higher reference likelihood therefore did not identify the model whose outputs were closest to the references.

cs.CL↗

Adapting LLM-SC to Living Sacred Heritage: Domain-Specific Extraction, Semantic Networks and Human Evaluation on Pashupatinath Temple Reviews

Large language model (LLM) pipelines can turn visitor reviews into structured data, but generic schemes may miss rituals, sacred spaces, cultural restrictions and emotionally mixed experiences at living sacred heritage sites. We adapted the LLM-Social Computing (LLM-SC) framework to Pashupatinath Temple, Nepal, by specifying a 12-type entity taxonomy (four original LLM-SC types plus eight sacred-heritage types), four relation types and nine aspect-sentiment dimensions in a GPT-4o extraction prompt, which was applied to a fixed random sample of 2,500 English-language Tripadvisor reviews from a processed corpus of 3,939. The exports contain 11,286 entity, 6,232 aspect-sentiment and 5,590 relation records, covering 2,485, 2,487 and 2,433 reviews, respectively; the eight added sacred-heritage types account for 64.8% of entity records. Against an adjudicated two-annotator reference for 200 reviews, exact normalized surface-and-type matching gave provisional entity scores of 41.0% precision, 21.8% recall and 28.4% micro-F1. Aspect-presence F1 was 53.6% and end-to-end aspect-plus-polarity F1 was 42.6%. Human adjudication supported 153 of 194 first-listed relation candidates (one per review; 78.9%), a candidate-level rate that does not estimate relation recall or whole-graph precision. The contribution is a culturally grounded, domain-specific adaptation with component-wise human evaluation and explicit reproducibility and generalization limitations.

cs.CL↗

Robust and Computationally Efficient Trimmed L-Moments Estimation for Parametric Distributions

This paper proposes a robust and computationally efficient estimation framework for fitting parametric distributions based on trimmed L-moments. Trimmed L-moments extend classical L-moment theory by downweighting or excluding extreme order statistics, resulting in estimators that are less sensitive to outliers and heavy tails. We construct estimators for both location-scale and shape parameters using asymmetric trimming schemes tailored to different moments, and establish their asymptotic properties for inferential justification using the general structural theory of L-statistics, deriving simplified single-integration expressions to ensure numerical stability. State-of-the-art algorithms are developed to resolve the sign ambiguity in estimating the scale parameter for location-scale models and the tail index for the Frechet model. The proposed estimators offer improved efficiency over traditional robust alternatives for selected asymmetric trimming configurations, while retaining closed-form expressions for a wide range of common distributions, facilitating fast and stable computation. Simulation studies demonstrate strong finite-sample performance. An application to financial claim severity modeling highlights the practical relevance and flexibility of the approach.

stat.ME↗