MASS-DPO derives a Plackett-Luce-specific log-determinant Fisher information objective to select non-redundant negative samples, matching or exceeding multi-negative DPO performance with substantially fewer negatives across four benchmarks and three model families.
Active preference optimization for sample efficient rlhf
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Positive-negative prompt pairing with weighted GRPO improves RLVR sample efficiency, raising AIME 2025 Pass@8 from 16.8 to 22.2 on Qwen2.5-Math-7B while matching large-scale training.
One training example via RLVR boosts LLM math reasoning from 17.6% to 35.7% average across six benchmarks.
Spectral Souping learns offline specialized policies for fine-grained preferences and merges them online using a discovered universal spectral representation for efficient LLM alignment.
DRRO for RLHF minimizes worst-case regret relative to the best policy under Wasserstein reward perturbations, yielding an exact inner solution and water-filling policy structure for the promptwise simplex model plus a practical policy-gradient algorithm.
PEPO is a single-step pessimistic ensemble algorithm for direct preference optimization that provably avoids over-optimization by depending only on single-policy concentrability without knowing the data distribution or learning an explicit reward model.
A statistical survey of RLHF for LLM alignment that connects preference learning and policy optimization to models like Bradley-Terry-Luce while reviewing methods, extensions, and open challenges.
citing papers explorer
-
MASS-DPO: Multi-negative Active Sample Selection for Direct Policy Optimization
MASS-DPO derives a Plackett-Luce-specific log-determinant Fisher information objective to select non-redundant negative samples, matching or exceeding multi-negative DPO performance with substantially fewer negatives across four benchmarks and three model families.
-
Beyond Variance: Prompt-Efficient RLVR via Rare-Event Amplification and Bidirectional Pairing
Positive-negative prompt pairing with weighted GRPO improves RLVR sample efficiency, raising AIME 2025 Pass@8 from 16.8 to 22.2 on Qwen2.5-Math-7B while matching large-scale training.
-
Reinforcement Learning for Reasoning in Large Language Models with One Training Example
One training example via RLVR boosts LLM math reasoning from 17.6% to 35.7% average across six benchmarks.
-
Spectral Souping: A Unified Framework for Online Preference Alignment
Spectral Souping learns offline specialized policies for fine-grained preferences and merges them online using a discovered universal spectral representation for efficient LLM alignment.
-
Wasserstein Distributionally Robust Regret Optimization for Reinforcement Learning from Human Feedback
DRRO for RLHF minimizes worst-case regret relative to the best policy under Wasserstein reward perturbations, yielding an exact inner solution and water-filling policy structure for the promptwise simplex model plus a practical policy-gradient algorithm.
-
Provably avoiding over-optimization in Direct Preference Optimization without knowing the data distribution
PEPO is a single-step pessimistic ensemble algorithm for direct preference optimization that provably avoids over-optimization by depending only on single-policy concentrability without knowing the data distribution or learning an explicit reward model.
-
Reinforcement Learning from Human Feedback: A Statistical Perspective
A statistical survey of RLHF for LLM alignment that connects preference learning and policy optimization to models like Bradley-Terry-Luce while reviewing methods, extensions, and open challenges.