Sample What You Say: Aligning Language Models to Sample the Distributions They State

arXiv:2609.34929 · cs.LG, cs.CL · Submitted 2026-09-28 · Read on arXiv

cs.LG, cs.CL

Submitted: 2026-09-28

Updated: 2026-09-28

Code: https://github.com/huggingface/trl

Terminology

• GRPO — 46× in this paper · explained in 6 episodes
• Alignment — 8× in this paper · explained in 1 episode
• In International Conference on Learning Representations — 8× in this paper
• Reinforcement Learning — 6× in this paper · explained in 4 episodes
• In Conference on Language Modeling — 5× in this paper
• Association for Computational Linguistics — 4× in this paper
• Benchmark — 4× in this paper · explained in 2 episodes
• Fine-Tuning — 4× in this paper · explained in 5 episodes
• Mode Collapse — 4× in this paper · explained in 2 episodes
• Distribution Matching — 3× in this paper · explained in 2 episodes
• Accuracy — 2× in this paper · explained in 1 episode
• Adam — 2× in this paper · explained in 2 episodes
• Calibration — 2× in this paper · explained in 7 episodes
• International Conference on Machine Learning — 2× in this paper
• Benchmarking — 1× in this paper · explained in 4 episodes
• Cosine Similarity — 1× in this paper · explained in 3 episodes
• Credit assignment — 1× in this paper · explained in 3 episodes
• Embeddings — 1× in this paper · explained in 2 episodes
• Empirical Methods in Natural Language Processing — 1× in this paper
• Geometric Distribution — 1× in this paper · explained in 2 episodes
• Group Relative Policy Optimization (GRPO) — 1× in this paper · explained in 5 episodes
• Hilbert Space — 1× in this paper · explained in 2 episodes
• Homogeneity — 1× in this paper · explained in 2 episodes
• Score Function — 1× in this paper · explained in 4 episodes
• Steering — 1× in this paper · explained in 2 episodes
• Transformer — 1× in this paper · explained in 2 episodes

Sources

• Back to Basics: Revisiting REINFORCE Style Optimization for Learning from Human Feedback in LLMs
• Group-Aware Reinforcement Learning for Output Diversity in Large Language Models
• Probabilistic Calibration Is a Trainable Capability in Language Models
• Demystifying MMD GANs
• The Entropy Mechanism of Reinforcement Learning for Reasoning Language Models
• Towards Measuring the Representation of Subjective Global Opinions in Language Models
• Differential Smoothing Mitigates Sharpening and Improves LLM Reasoning
• The Llama 3 Herd of Models
• The Illusion of Stochasticity in LLMs
• Rewarding the Unlikely: Lifting GRPO Beyond Distribution Sharpening
• Amortizing intractable inference in large language models
• Self-Improvement in Language Models: The Sharpening Mechanism
• Alignment Defends LLMs from Property Inference Attacks
• Instruction-Tuned Language Models Cannot Sample from Distributions They Can Describe
• Artificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond)
• A Distributional Approach to Controlled Text Generation
• Understanding the Effects of RLHF on LLM Generalisation and Diversity
• The Choice of Divergence: A Neglected Key to Mitigating Diversity Collapse in Reinforcement Learning with Verifiable Reward
• Jointly Reinforcing Diversity and Quality in Language Model Generations
• Beyond Mode Collapse: Distribution Matching for Diverse Reasoning

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks