Premise-Augmented Reasoning Chains Improve Error Identification in Math reasoning with LLMs