Identifying Interactions at Scale for LLMs
BAIR 7 months ago 39
Researchers developed SPEX and ProxySPEX, algorithms that efficiently identify influential feature interactions in large language models by leveraging sparsity and hierarchical properties to reduce computational costs from exponential to tractable levels. ProxySPEX achieves the same performance as SPEX with approximately 10 times fewer ablations required. The methods enable new applications in feature attribution, data attribution, and mechanistic interpretability across different scales of model analysis, with code made available in the SHAP-IQ repository.