Stable Alignment

Liu and coauthors use a simulated society of language-model agents to generate ratings, feedback, and revised responses for training. They call their framework Stable Alignment.^[1]

The paper reports improvements on its alignment evaluations. Simulated agreement does not by itself establish that real communities endorse the resulting norms. The paper writeup describes the training stages and limits.^[1]

Ruibo Liu and coauthors explicitly limit their reported results to primarily English, text-based settings and warn that simulated feedback can miss underrepresented communities.^[1] Preference Aggregation addresses a related but different problem: combining preferences in a decision rule. Who Supplies Alignment Feedback? separates feedback availability from reliability and representation, connecting this method to the wider question of Social Alignment.