Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging

TL;DR

This paper introduces a simple post-hoc model merging technique that navigates the alignment-calibration trade-off, producing models that outperform their parents in accuracy and calibration, thus mitigating the alignment tax efficiently.

Contribution

It demonstrates that interpolating between pre- and post-alignment models reveals Pareto-optimal solutions, improving both accuracy and calibration beyond individual models.

Findings

Interpolating models recovers calibration lost during alignment.

Model merging reveals Pareto-optimal trade-offs.

Merged models outperform individual models in accuracy.

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.