Bridging the Reasoning Gap in Vietnamese with Small Language Models via Test-Time Scaling

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Trung, Bui The, Duc, Do Minh, Van Vinh, Nguyen, Trinh, Bui Nguyen Quoc
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908978576883712
author Trung, Bui The
Duc, Do Minh
Van Vinh, Nguyen
Trinh, Bui Nguyen Quoc
author_facet Trung, Bui The
Duc, Do Minh
Van Vinh, Nguyen
Trinh, Bui Nguyen Quoc
contents The democratization of ubiquitous AI hinges on deploying sophisticated reasoning capabilities on resource-constrained devices. However, Small Language Models (SLMs) often face a "reasoning gap", particularly in non-English languages like Vietnamese, where they struggle to maintain coherent chains of thought. This paper investigates Test-Time Scaling strategies for the Qwen3-1.7B architecture within the context of Vietnamese Elementary Mathematics. We introduce Vi-S1K, a high-fidelity reasoning dataset localized via a Gemini 2.5 Flash-Lite powered pipeline, and Vi-Elementary-Bench, a dual-resource benchmark for rigorous evaluation. Using an LLM-as-a-Judge protocol, we reveal that the base model possesses robust latent knowledge (Accuracy: 4.05/5.00) but suffers from a severe "formatting gap" in communication. Supervised Fine-Tuning (SFT) acts as a critical "reasoning unlocker", yielding a 77% improvement in Explanation Quality and bridging the gap between raw calculation and pedagogical coherence. Furthermore, our analysis of prompting strategies uncovers a significant trade-off: structured frameworks like ReAct impose a "cognitive tax" on the 1.7B parameter capacity, degrading performance relative to pure Chain-of-Thought (CoT) combined with Self-Consistency. These findings establish a deployment hierarchy for SLMs, demonstrating that SFT combined with simplified test-time scaling is superior to complex agentic workflows for edge-based reasoning.
format Preprint
id arxiv_https___arxiv_org_abs_2604_17794
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Bridging the Reasoning Gap in Vietnamese with Small Language Models via Test-Time Scaling
Trung, Bui The
Duc, Do Minh
Van Vinh, Nguyen
Trinh, Bui Nguyen Quoc
Computation and Language
Artificial Intelligence
The democratization of ubiquitous AI hinges on deploying sophisticated reasoning capabilities on resource-constrained devices. However, Small Language Models (SLMs) often face a "reasoning gap", particularly in non-English languages like Vietnamese, where they struggle to maintain coherent chains of thought. This paper investigates Test-Time Scaling strategies for the Qwen3-1.7B architecture within the context of Vietnamese Elementary Mathematics. We introduce Vi-S1K, a high-fidelity reasoning dataset localized via a Gemini 2.5 Flash-Lite powered pipeline, and Vi-Elementary-Bench, a dual-resource benchmark for rigorous evaluation. Using an LLM-as-a-Judge protocol, we reveal that the base model possesses robust latent knowledge (Accuracy: 4.05/5.00) but suffers from a severe "formatting gap" in communication. Supervised Fine-Tuning (SFT) acts as a critical "reasoning unlocker", yielding a 77% improvement in Explanation Quality and bridging the gap between raw calculation and pedagogical coherence. Furthermore, our analysis of prompting strategies uncovers a significant trade-off: structured frameworks like ReAct impose a "cognitive tax" on the 1.7B parameter capacity, degrading performance relative to pure Chain-of-Thought (CoT) combined with Self-Consistency. These findings establish a deployment hierarchy for SLMs, demonstrating that SFT combined with simplified test-time scaling is superior to complex agentic workflows for edge-based reasoning.
title Bridging the Reasoning Gap in Vietnamese with Small Language Models via Test-Time Scaling
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2604.17794