Fast XAI with interactions at large scale. SPEX can help you understand the output of your LLM, even if you have a long context!
