This table analyzes output-space LLM merging approaches that combine models through response-level aggregation instead of parameter manipulation. It details their fusion mechanisms, architectural cons

DataCite Commons2026-03-02 更新00

Comparison of Output-Space LLM Merging Methods

This table compares output-space LLM merging methods, where models are combined at the level of generated responses rather than internal parameters. It analyzes their usion strategies, architectural a

DataCite Commons2026-02-20 更新10

LLM Merging Methods in the weight space

This table analyzes weight-space LLM merging approaches that operate directly on model parameters. It characterizes their architectural constraints, training requirements, computational cost, experime

DataCite Commons2026-03-02 更新10

LLM Merging Methods in the output space