The paper introduces OGC-DPO, an active preference learning framework for aligning large language models with multiple objectives such as helpfulness, harmlessness, and honesty. OGC-DPO treats each response pair and objective as a separate query, using the model’s implicit gradient to gauge informativeness and selecting queries that expand gradient coverage for each objective. Experiments show that OGC-DPO outperforms existing methods at the same annotation cost and reveals that informative conflicts can serve as valuable training signals.
Aligning large language models with diverse and heterogeneous human values requires multi-objective alignment methods to effectively trade off conflicting preference dimensions. Current methods achieve this trade-off by training policies conditioned on preference vectors and leveraging online direct preference optimization.
arXiv:2512.23067v3 Announce Type: replace Abstract: Personalized alignment from preference data has focused primarily on improving personal reward model (RM) accuracy, with the implicit assumption th...
UniPolicy is a unified objective‑specific policy framework for search advertising that jointly optimizes relevance, click propensity, and commercial value. It uses objective‑aware prefix tokens, sparse MoE‑LoRA routing, and residual FFNs to decouple parameters within a shared backbone, and constructs pairwise preferences from multi‑stage behavioral feedback to strengthen clicked candidates. In large‑scale offline tests and a 7‑day online A/B test, UniPolicy improves CTR by 0.71%, RPS by 1.58%, and advertising revenue by 1.32% while keeping serving latency stable.
UniPolicy is a multi-policy alignment framework for search advertising that jointly optimizes relevance, click propensity, and commercial value. It uses objective-specific prefix tokens, sparse MoE-LoRA routing, and residual FFNs to decouple parameters within a shared backbone, and builds pairwise preferences from multi-stage behavioral feedback to improve generation. In large-scale offline tests and a 7‑day online A/B test, UniPolicy achieved balanced gains across metrics, boosting CTR by 0.71%, RPS by 1.58%, and revenue by 1.32% while keeping latency stable.
arXiv:2609.38860v1 Announce Type: cross Abstract: Learning from human preferences is central to large language model (LLM) alignment, but human preference annotation is costly. Active preference lear...
arXiv:2504. 06659v2 Announce Type: replace-cross Abstract: Despite advances in Preference Alignment (PA) for Large Language Models (LLMs), mainstream methods like reinforcement learning with human feedback face notable challenges.
arXiv:2504.20106v4 Announce Type: replace Abstract: Ensuring that large language models (LLMs) are both helpful and harmless is a critical challenge, as overly strict constraints can lead to excessiv...
arXiv:2510. 11194v3 Announce Type: replace Abstract: Personalized alignment is crucial for enabling Large Language Models (LLMs) to engage effectively in user-centric interactions.
The paper introduces Collaborative Reward Modeling (CRM), an online framework that uses cross-model peer review and curriculum learning to identify reliable preference pairs for training reward models (RMs). By having two RMs evaluate each other's sample selections and synchronizing their learning progress, CRM reduces the impact of noisy feedback, improving RM generalization by up to 9.94 points on RewardBench under 40% preference noise. The approach also extends to implicit-reward alignment, demonstrating robustness and broad applicability for learning from imperfect feedback.
arXiv:2606. 04807v1 Announce Type: new Abstract: Mitigating social bias in Large Language Models (LLMs) presents a distinct alignment challenge: unlike verifiable tasks, bias lacks a single ground truth, creating a high-variance, subjective reward landscape.
arXiv:2609.08211v1 Announce Type: new Abstract: Real-world Multi-Objective Reinforcement Learning (MORL) often suffers from sparse rewards, reward conflicts, and late-stage reward tug-of-war, causing...
The paper reproduces five factuality alignment methods for large language models—three on‑policy RL approaches (PPO, GRPO, Dr.GRPO) and two offline preference optimization methods (DPO, KTO)—using a common benchmark and evaluation metrics. None of the methods consistently outperforms the base model, and the rewards appear to favor format, length, or brevity rather than factuality or relevance. The study also highlights problems with current evaluation practices, noting that metrics and benchmarks often fail to measure prompt relevance and rely on fragile LLM judges and retrieval for verification.