Unconstrained Model Merging for Enhanced LLM Reasoning

Recent advancements in building domain-specific large language models (LLMs) have shown remarkable success, especially in tasks requiring reasoning abilities like logical inference over complex relationships and multi-step problem solving. However, creating a powerful all-in-one LLM remains challenging due to the need for proprietary data and vast computational resources. As a resource-friendly alternative, we explore the potential of merging multiple expert models into a single LLM. Existing studies on model merging mainly focus on generalist LLMs instead of domain experts, or the LLMs under the same architecture and size. In this work, we propose an unconstrained model merging framework that accommodates both homogeneous and heterogeneous model architectures with a focus on reasoning tasks. A fine-grained layer-wise weight merging strategy is designed for homogeneous models merging, while heterogeneous model merging is built upon the probabilistic distribution knowledge derived from instruction-response fine-tuning data. Across 7 benchmarks and 9 reasoning-optimized LLMs, we reveal key findings that combinatorial reasoning emerges from merging which surpasses simple additive effects. We propose that unconstrained model merging could serve as a foundation for decentralized LLMs, marking a notable progression from the existing centralized LLM framework. This evolution could enhance wider participation and stimulate additional advancement in the field of artificial intelligence, effectively addressing the constraints posed by centralized models.

NeuronBlocks -- Building Your NLP DNN Models Like Playing Lego

When building deep neural network models for natural language processing tasks, engineers often spend a lot of efforts on coding details and debugging, instead of focusing on model architecture design and hyper-parameter tuning. In this paper, we introduce NeuronBlocks, a deep neural network toolkit for natural language processing tasks. In NeuronBlocks, a suite of neural network layers are encapsulated as building blocks, which can easily be used to build complicated deep neural network models by configuring a simple JSON file. NeuronBlocks empowers engineers to build and train various NLP models in seconds even without a single line of code. A series of experiments on real NLP datasets such as GLUE and WikiQA have been conducted, which demonstrates the effectiveness of NeuronBlocks.