An Empirical Study of Sustainability in Prompt-driven Test Script Generation Using Small Language Models

arXiv

View on arXiv

2.0/10

Platform Domination RiskN/A

Market Consolidation RiskN/A

Displacement HorizonN/A

CORE FUNCTION

Empirical analysis and benchmarking of small language models (2B-8B) for unit-test generation, specifically measuring energy consumption and carbon emissions alongside code quality.

TRACTION

citations

0.0 velocity

co_authors

0.0 velocity

REASONING

The project is primarily a research study (paper-based) rather than a software tool. While it addresses a niche gap—sustainability of SLMs in test generation—it lacks a functional codebase with adoption or a technical moat. Its value is purely academic and data-driven.

COMPOSABILITY

TECH STACK

PythonPyTorchTransformersSmall Language Models (SLMs)Energy measurement tools (e.g., CodeCarbon)JUnit/Unit testing frameworks

INTEGRATION

reference_implementation

sustainability_benchmarkingautomated_test_generationslm_evaluationcarbon_footprint_analysis

READINESS

Composabilitytheoretical

Depth