Go to ICLR 2026 Workshop Trustworthy AI homepageThe Realignment Problem: When Right becomes Wrong in LLMs
Keywords: Preference Optimization, Machine Unlearning, Reinforcement Learning
TL;DR: We re-align LLMs to new values without re-annotation by programmatically identifying and correcting policy conflicts in existing preference data.
Abstract: Post-training alignment of large language models (LLMs) relies on large-scale human annotations guided by policy specifications that change over time. Cultural shifts, value reinterpretations, and regulatory or industrial updates make static alignment increasingly brittle. As policies evolve, deployed models can diverge from current alignment objectives, creating an Alignment–Reality Gap that is difficult to audit or correct. Existing remediation typically requires re-annotation under revised guidelines, which introduces systematic challenges, including guideline ambiguity, annotator interpretation drift, and reduced consistency at scale. We introduce TRACE (Triage and Re-align by Alignment Conflict Evaluation), a framework that transforms re-alignment into a structured optimization problem over existing data without requiring fresh human annotation. Leveraging a stronger model as a proxy judge, TRACE operates via a three-stage pipeline: (1) triaging preference pairs into inversion, suppression, or retention categories based on alignment conflicts; (2) computing an alignment impact score via bi-level optimization to prioritize high-leverage samples; and (3) executing updates using a hybrid objective that combines relational losses (e.g., IPO) for preference inversion and punitive losses (e.g., NPO) for response suppression. Experiments on Qwen2.5-7B, Gemma-2-9B, and Llama-3.1-8B demonstrate robust re-alignment on synthetic benchmarks and the PKU-SafeRLHF dataset without degrading general utility. This work provides a scalable approach for LLM realignment under evolving data annotation policies and alignment guidelines.
Submission Number: 259
Loading