arXiv AI
Jul 21
Normalized Rewards for Preference Optimization
arXiv:2607. 16240v1 Announce Type: cross Abstract: Direct Alignment Algorithms (DAAs) such as DPO have become a common way to post-train and align LLMs with human preferences.
MINT: Min-Selection Preference Distillation for Balanced Multi-Objective Alignment
arXiv:2608. 14828v1 Announce Type: new Abstract: Aligning a language agent to several objectives at once is a persistent failure mode of preference-based training: when objectives are combined additively, optimization collapses onto whichever is cheapest to improve and sacrifices the rest, so a support agent learns to sound warm while giving no real help.
Related stories
arXiv AI
Jul 24
Towards Disentangled Preference Optimization Dynamics: Suppress the Loser, Preserve the Winner
arXiv:2604. 18239v4 Announce Type: replace-cross Abstract: Preference optimization is widely used to align large language models (LLMs) with human preferences.
arXiv AI
Sep 1
Personalized Group Relative Policy Optimization for Heterogenous Preference Alignment
The paper introduces Personalized Group Relative Policy Optimization (P‑GRPO), a new alignment framework for large language models that separates advantage estimation from immediate batch statistics. By normalizing advantages using preference‑group‑specific reward histories instead of the concurrent generation group, P‑GRPO maintains contrastive signals for distinct user preferences. Experiments across various tasks show that P‑GRPO converges faster and yields higher rewards than standard GRPO, improving alignment with heterogeneous human preferences while preserving general capabilities.
arXiv Machine Learning
Jun 2
The Representation-Rationalizability Tradeoff in Reward Learning
arXiv:2606. 00291v1 Announce Type: cross Abstract: In RLHF, each training example contains a prompt $x$ and two candidate responses $y,y'$, and annotators provide pairwise preferences between these responses.
arXiv Machine Learning
Sep 22
Swiss-Knife: A Framework for Reconfigurable Externalised Multi-Objective Alignment at Decode Time
Swiss-Knife is a framework that extends decode‑time alignment for frozen language models by treating the alignment specification as a runtime object. It introduces hot‑swappable scoring blades, a batch normaliser, a pairwise aggregation operator, and a selection rule, and characterises admissible aggregation operators with a representation theorem. In experiments, Swiss‑Knife paired with DPO‑LoRA blades and an uncertainty‑aware pairwise tournament outperforms six existing decode‑time methods, achieving a higher harmonic F1 score, lower refusal rate, and faster objective reconfiguration.
arXiv Machine Learning
Jun 16
GD$^2$PO: Mitigating Multi-Reward Conflicts via Group-Dynamic reward-Decoupled Policy Optimization
arXiv:2606. 16771v1 Announce Type: new Abstract: As LLMs advance, post-training reinforcement learning (RL) increasingly relies on multi-dimensional rewards to cultivate comprehensive capabilities.
arXiv AI
Aug 12
Multiplayer Nash Preference Optimization
arXiv:2509. 23102v4 Announce Type: replace Abstract: Reinforcement learning from human feedback (RLHF) has emerged as the standard paradigm for aligning large language models with human preferences.
arXiv AI
Jun 2
Margin Adaptive DPO: Leveraging Reward Model for Granular Control in Preference Optimization
arXiv:2510. 05342v2 Announce Type: replace-cross Abstract: Direct Preference Optimization (DPO) has emerged as a simple and effective method for aligning large language models.
arXiv AI
Aug 12
Procedural Fairness Failures in RLHF from Preference Averaging
arXiv:2608. 10126v1 Announce Type: cross Abstract: Reinforcement Learning from Human Feedback (RLHF) aggregates heterogeneous preferences into a single reward model, assuming preference homogeneity.
arXiv Machine Learning
Sep 22
Gradient-Gated DPO: Stabilizing Preference Optimization in Language Models
arXiv:2605.02626v2 Announce Type: replace Abstract: Direct Preference Optimization (DPO) improves relative preference by increasing the margin between chosen and rejected responses, but this objectiv...
arXiv AI
Jul 7
Out-of-Distribution Generalization of Risk Aversion in Language Models
arXiv:2607. 02755v1 Announce Type: cross Abstract: Training AIs to be risk-averse in resources could offer a failsafe in the event that AIs turn out misaligned.
arXiv AI
Jul 7
Attention Limited Reward Learning
arXiv:2607. 04590v1 Announce Type: new Abstract: Pairwise human comparisons are a primary interface through which modern AI systems learn human preferences.