On the Ability of LLMs to Handle Character-Level Perturbations: How Well and How?

Open in new window