MSDDA derives a closed-form optimal reverse denoising distribution for multi-objective diffusion alignment that is exactly equivalent to step-level RL fine-tuning with no approximation error.
Rewarded soups: towards pareto-optimal alignment by interpolating weights fine-tuned on diverse rewards.Advances in Neural Information Processing Systems, 36:71095–71134
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
method 1
citation-polarity summary
years
2026 4verdicts
UNVERDICTED 4roles
method 1polarities
use method 1representative citing papers
SURF derives weight sampling rules from the arc-length CDF of the scalarization path to uniformly traverse the Pareto front in multi-objective optimization.
GPRL carries a k-dimensional skew-symmetric preference structure into policy updates with per-dimension advantages and a drift monitor, yielding 56.51% length-controlled win rate on AlpacaEval 2.0 from Llama-3-8B-Instruct while outperforming SimPO and SPPO on other benchmarks.
MORA breaks the safety-helpfulness ceiling in LLMs by pre-sampling single-reward prompts and rewriting them to incorporate multi-dimensional intents, delivering 5-12.4% gains in sequential alignment and 4.6% overall improvement in simultaneous alignment.
citing papers explorer
-
Step-level Denoising-time Diffusion Alignment with Multiple Objectives
MSDDA derives a closed-form optimal reverse denoising distribution for multi-objective diffusion alignment that is exactly equivalent to step-level RL fine-tuning with no approximation error.
-
SURF: Steering the Scalarization Weight to Uniformly Traverse the Pareto Front
SURF derives weight sampling rules from the arc-length CDF of the scalarization path to uniformly traverse the Pareto front in multi-objective optimization.
-
General Preference Reinforcement Learning
GPRL carries a k-dimensional skew-symmetric preference structure into policy updates with per-dimension advantages and a drift monitor, yielding 56.51% length-controlled win rate on AlpacaEval 2.0 from Llama-3-8B-Instruct while outperforming SimPO and SPPO on other benchmarks.
-
Explaining and Breaking the Safety-Helpfulness Ceiling via Preference Dimensional Expansion
MORA breaks the safety-helpfulness ceiling in LLMs by pre-sampling single-reward prompts and rewriting them to incorporate multi-dimensional intents, delivering 5-12.4% gains in sequential alignment and 4.6% overall improvement in simultaneous alignment.