Add new paper: Interpreting Arithmetic Mechanism in Large Language Models through Comparative Neuron Analysis

[fan2goa1]

Title: Interpreting Arithmetic Mechanism in Large Language Models through Comparative Neuron Analysis Head: Arithmetic Head Published: EMNLP 2024 Summary:

• Innovation: Introduces Comparative Neuron Analysis (CNA) to map arithmetic mechanisms in attention heads of large language models.
• Tasks: Analyzed arithmetic ability, model pruning for arithmetic tasks, and model editing to reduce gender bias.
• Significant Result: Identified specific neurons responsible for arithmetic, enabling performance improvements and bias mitigation through targeted neuron manipulation.