Follow
Frontier Language Models are not Robust to Adversarial Arithmetic, or "What do I need to say so you agree 2+2=5?. (arXiv:2311.07587v2 [cs.CL] UPDATED)
Frontier Language Models are not Robust to Adversarial Arithmetic, or "What do I need to say so you agree 2+2=5?. (arXiv:2311.07587v2 [cs.CL] UPDATED)