arXiv Artificial Intelligence

Procedural Fairness Failures in RLHF from Preference Averaging

1 day ago 12.08.2026 07:00 👁 2 views
Procedural Fairness Failures in RLHF from Preference Averaging

Quick summary

arXiv:2608.10126v1 Announce Type: cross Abstract: Reinforcement Learning from Human Feedback (RLHF) aggregates heterogeneous preferences into a single reward model, assuming preference homogeneity. When preferences are heterogeneous, this aggregation induces a procedural fairness failure where majority preference groups dominate reward learning while minority preferences are systematically under-represented. This work defines procedural fairness in alignment as preserving distinct preference signals during reward modeling and shows that standard RLHF violates this via preference averaging. Pre

Key takeaways

  • arXiv:2608.10126v1 Announce Type: cross Abstract: Reinforcement Learning from Human Feedback (RLHF) aggregates heterogeneous preferences into a single reward model, assuming preference homogeneity.
  • When preferences are heterogeneous, this aggregation induces a procedural fairness failure where majority preference groups dominate reward learning while minority preferences are systematically under-represented.
  • This work defines procedural fairness in alignment as preserving distinct preference signals during reward modeling and shows that standard RLHF violates this via preference averaging.

Why it matters

“Procedural Fairness Failures in RLHF from Preference Averaging” should be evaluated beyond branding and benchmark scores. Its practical importance will emerge in task accuracy, latency, unit cost, safety and integration with real workflows.

Kaynak sitede devamını oku: arXiv Artificial Intelligence ↗