LLM Merging methods for reasoning Tasks
Abstract
This table analyzes LLM merging approaches targeting reasoning tasks, including benchmarks such as GSM8K, Common-sense and MATH. It examines their fusion mechanisms, architectural assumptions, training dependencies, and computational cost, clarifying how different merging paradigms impact multi-step logical and quantitative reasoning performance.
Fields of Study & Subjects
Topic Modeling
Multimodal Machine Learning Applications
Domain Adaptation and Few-Shot Learning
Publication details
- Journal / source
- Open Research Knowledge Graph (ORKG)
- Publication date
- 3 March 2026
- Language
- EN
- Document type
- Dataset