Search papers, labs, and topics across Lattice.
This paper introduces a federated reasoning distillation framework to train a small, efficient financial reasoning expert LM by leveraging Chain-of-Thought (CoT) reasoning from multiple teacher LLMs. The framework refines CoTs from various teacher LLMs using an "LLM-as-a-Judge" approach to improve the supervision signal for the student model. Experiments demonstrate that a 7B distilled LM achieves competitive performance on financial question-answering datasets by learning from these refined CoTs.
A 7B language model, distilled from multiple larger LLMs using a federated Chain-of-Thought approach, can achieve surprisingly strong performance on financial reasoning tasks.
Recent advancements in Large Language Models (LLMs) have significantly enhanced reasoning capabilities for complex tasks. However, the substantial computational demands and high deployment costs of these LLMs remain significant barriers to their widespread adoption. Although progress has been made in applying knowledge distillation to improve the performance of resource-friendly small LMs on general tasks, developing compact yet powerful financial reasoning expert LMs remains underexplored. In this paper, we present a comprehensive evaluation of leading reasoning and non-reasoning LLMs on their financial reasoning abilities by challenging these models to solve financial problems with detailed reasoning Chain-of-Thought (CoTs). Based on the benchmark results, we introduce a federated reasoning distillation framework where financial reasoning CoTs from various teacher LLMs are refined following "LLM-as-a-Judge" fashion to better supervise the training process of the student financial expert. Our experiments show that a 7B distilled LM can achieve competitive performance on two public finance question-answer datasets by learning from multiple teachers. To foster future research in financial AI, we will open-source the refined financial CoT dataset at https://github.com/yjump/FinCoT.
