{"ID":2852716,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.17426","arxiv_id":"2510.17426","title":"Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging","abstract":"The \"alignment tax\" of post-training is typically framed as a drop in task accuracy. We show it also involves a severe loss of calibration, making models overconfident, less reliable, and model outputs less diverse. We show that this trade-off can be navigated effectively via a simple post-hoc intervention: interpolating between a model's weights before and after alignment. Crucially, this is not a strict trade-off. We find that the process consistently reveals Pareto-optimal interpolations - models that improve accuracy beyond both parents while substantially recovering the calibration lost during alignment. Our work demonstrates that simple model merging provides a computationally efficient method for mitigating the full scope of the alignment tax, yielding models that are more capable and more reliable.","short_abstract":"The \"alignment tax\" of post-training is typically framed as a drop in task accuracy. We show it also involves a severe loss of calibration, making models overconfident, less reliable, and model outputs less diverse. We show that this trade-off can be navigated effectively via a simple post-hoc intervention: interpolati...","url_abs":"https://arxiv.org/abs/2510.17426","url_pdf":"https://arxiv.org/pdf/2510.17426v2","authors":"[\"Tiancheng Hu\",\"Benjamin Minixhofer\",\"Nigel Collier\"]","published":"2025-10-20T11:12:41Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\",\"cs.LG\"]","methods":"[]","has_code":false}