Zero-RL multi-stage constructive safety alignment with SERL and long-context training lets a 14B model match much larger models on safety without collapsing helpfulness or style.
Orpo: Monolithic preference optimization without reference model
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
citation-role summary
background 1
method 1
citation-polarity summary
representative citing papers
Bidirectional LLM-GNN co-teaching with round-based pseudo-label preference optimization outperforms golden-teacher baselines on few-shot TAG benchmarks by 3-8% absolute gains.
Super-Linear Advantage Shaping (SLAS) introduces a non-linear geometric policy update for RL post-training of text-to-image models that reshapes the local policy space via advantage-dependent Fisher-Rao weighting to reduce reward hacking and improve performance over GRPO baselines.
A preference-learning framework that augments RLHF with intra-sample and inter-sample prompt-level objectives, improving multi-instruction following without lowering semantic quality on tested benchmarks.
LeanPO improves Video-LLM alignment by using a reference-free average-likelihood reward, self-generated winning/losing pairs, and dynamic label smoothing, yielding gains on six video benchmarks.
citing papers explorer
-
Oyster-II: Reinforcement Learning for Constructive Safety Alignment in Large Language Models
Zero-RL multi-stage constructive safety alignment with SERL and long-context training lets a 14B model match much larger models on safety without collapsing helpfulness or style.
-
Beyond the Golden Teacher: Enhancing Graph Learning through LLM-GNN Co-teaching
Bidirectional LLM-GNN co-teaching with round-based pseudo-label preference optimization outperforms golden-teacher baselines on few-shot TAG benchmarks by 3-8% absolute gains.
-
Power Reinforcement Post-Training of Text-to-Image Models with Super-Linear Advantage Shaping
Super-Linear Advantage Shaping (SLAS) introduces a non-linear geometric policy update for RL post-training of text-to-image models that reshapes the local policy space via advantage-dependent Fisher-Rao weighting to reduce reward hacking and improve performance over GRPO baselines.
-
Multi-Level Aware Preference Learning: Enhancing RLHF for Complex Multi-Instruction Tasks
A preference-learning framework that augments RLHF with intra-sample and inter-sample prompt-level objectives, improving multi-instruction following without lowering semantic quality on tested benchmarks.
-
LeanPO: Lean Preference Optimization for Likelihood Alignment in Video-LLMs
LeanPO improves Video-LLM alignment by using a reference-free average-likelihood reward, self-generated winning/losing pairs, and dynamic label smoothing, yielding gains on six video benchmarks.
- Beyond Pairs: Your Language Model is Secretly Optimizing a Preference Graph