Logistic Regression makes small LLMs strong and explainable "tens-of-shot" classifiers

Re-assign community