Accelerating Greedy Coordinate Gradient and General Prompt Optimization via Probe Sampling

Open in new window