Challenge: Existing approaches to align large language models with human preferences are limited in generalizability due to distribution shift, preference label noise, and mismatch of challenging samples with model capacity.
Approach: They propose a framework that constructs preference pairs with varying difficulty levels and then produces a specific curriculum for reward model training.
Outcome: The proposed framework improves generalizability of reward models by a significant margin without incurring additional inference costs compared to existing non-curriculum baselines.

Similar Papers

Reward Generalization in RLHF: A Topological Perspective (2025.findings-acl)

Copied to clipboard

Challenge: Existing alignment methods share a common topology of information flow, but their alternatives have not been thoroughly explored.
Approach: They propose a theory of reward generalization in reinforcement learning from human feedback . they propose induced Bayesian networks to model the impact of dataset topologies on reward generalisation .
Outcome: The proposed method achieves an average win rate of 65% on three NLP tasks.
Fine-Tuning Language Models with Reward Learning on Policy (2024.naacl-long)

Copied to clipboard

Challenge: Reinforcement learning from human feedback (RLHF) is an effective approach to align large language models (LLMs) to human preferences.
Approach: They propose a framework that refines a reward model using policy samples to keep it on-distribution.
Outcome: The proposed framework outperforms the state-of-the-art on three benchmark datasets showing that it can learn robust representations of policy samples.
Towards Reward Fairness in RLHF: From a Resource Allocation Perspective (2025.acl-long)

Copied to clipboard

Challenge: if rewards are imperfect, they can adversely affect the alignment of large language models (LLMs).
Approach: They propose a bias-agnostic method to address the issue of reward unfairness from a resource allocation perspective without specifically designing for each type of bias . they apply methods Fairness Regularization and Fairness Coefficient to achieve fairness in rewards.
Outcome: The proposed method achieves fairness in rewards while minimizing biases . it can be applied to verification and reinforcement learning scenarios .
PARIF: Pushing the Pareto Frontier of Instruction Following and Reasoning with Curriculum Reinforcement Learning (2026.acl-long)

Copied to clipboard

Challenge: Existing alignment methods struggle to balance general reasoning with instruction-following (IF) this is hindered by dependency on teacher models, reward hacking, and reasoning-answer inconsistencies.
Approach: They propose a two-stage curriculum learning framework based on Reinforcement Learning from Verifiable Rewards to enhance both IF and general reasoning capabilities.
Outcome: The proposed framework outperforms leading models on six representative IF tasks while achieving a 21.25% relative average improvement over the original model.
RED: Unleashing Token-Level Rewards from Holistic Feedback via Reward Redistribution (2025.emnlp-main)

Copied to clipboard

Challenge: Experimental results demonstrate the superiority of our approach to aligning large language models with human preferences.
Approach: They propose a method that evaluates and assigns specific credit to each token using an off-the-shelf reward model.
Outcome: The proposed method evaluates and assigns specific credit to each token using an off-the-shelf reward model.
Don’t Forget Your Reward Values: Language Model Alignment via Value-based Calibration (2024.emnlp-main)

Copied to clipboard

Challenge: Existing methods for generating large language models have been criticized for their complexity and instability.
Approach: They propose a value-based calibration method to better align Large Language Models with human preferences.
Outcome: The proposed method surpasses existing methods on AI assistant and summarization datasets, providing impressive generalizability, robustness, and diversity in different settings.
Reward Modeling Requires Automatic Adjustment Based on Data Quality (2024.findings-emnlp)

Copied to clipboard

Challenge: Reinforcement Learning from Human Feedback (RLHF) is a method for aligning language models with human values.
Approach: They propose a method that automatically adjusts reward modeling based on data quality . they use preference data to train a reward model that is more aligned with human values .
Outcome: The proposed method stabilizes reward model training and significantly improves alignment performance on human preference datasets.
On the Limited Generalization Capability of the Implicit Reward Model Induced by Direct Preference Optimization (2024.findings-emnlp)

Copied to clipboard

Challenge: Reinforcement Learning from Human Feedback (RLHF) is an effective approach for aligning language models to human preferences.
Approach: They compare the accuracy of DPORM and EXRM with a reward function for scoring human preferences.
Outcome: The proposed methods can approximate an EXRM on the limit infinite samples, but it is unclear how effective they are in practice.
Prototypical Reward Network for Data-Efficient Model Alignment (2024.acl-long)

Copied to clipboard

Challenge: Reinforcement Learning from Human Feedback (RLHF) is a reward model that fine-tunes Large Language Models (LLMs) by utilizing Prototypical Networks.
Approach: They propose a framework utilizing Prototypical Networks to enhance reward models under limited human feedback, enabling more stable and reliable structural learning from fewer samples.
Outcome: The proposed framework improves reward models under limited human feedback, surpassing traditional methods, especially in data-limited scenarios.
LIRE: listwise reward enhancement for preference alignment (2024.findings-acl)

Copied to clipboard

Challenge: prevailing approaches to preference alignment focus on pairwise comparisons, with limited exploration into multi-response scenarios.
Approach: They propose a listwise reward enhancement approach that integrates offline rewards of multiple responses into a streamlined listwise framework.
Outcome: The proposed approach outperforms existing methods on dialogue and summarization tasks with good transferability to out-of-distribution data.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations