Apr 20
LLM attribution analysis across different fine-tuning strategies and model scales for automated code compliance
★★★★★
significance 2/5
This research investigates how different fine-tuning strategies and model scales affect the interpretability of LLMs in automated code compliance. The study uses perturbation-based attribution analysis to show that full fine-tuning produces more focused attribution patterns compared to parameter-efficient methods like LoRA.
Why it matters
Fine-tuning depth directly dictates the precision of model reasoning, a critical factor for high-stakes automated code compliance and auditing.
Tags
#llm #fine-tuning #interpretability #code compliance #attributionRelated coverage
- Global South OpportunitiesPivotal Research Fellowship 2026 (Q3): AI Safety Research Opportunity - Global South Opportunities
- arXiv cs.AIAn Intelligent Fault Diagnosis Method for General Aviation Aircraft Based on Multi-Fidelity Digital Twin and FMEA Knowledge Enhancement
- arXiv cs.AIPExA: Parallel Exploration Agent for Complex Text-to-SQL
- arXiv cs.AIThe Power of Power Law: Asymmetry Enables Compositional Reasoning
- arXiv cs.AIOn the Existence of an Inverse Solution for Preference-Based Reductions in Argumentation