AINeutralarXiv – CS AI · Jun 96/10
🧠A position paper argues that large language models should optimize for individual user preferences rather than aggregated 'average user' preferences, which masks critical information about preference diversity and values. The authors propose bounded personalization frameworks that balance individual autonomy with universal safety constraints, while addressing scalability and manipulation risks.
AIBullisharXiv – CS AI · Jun 46/10
🧠Researchers introduce BiasGRPO, a novel framework using Group Relative Policy Optimization to mitigate social bias in Large Language Models more effectively than existing methods. The approach stabilizes training in high-variance reward landscapes by normalizing rewards across sampled completions, outperforming Direct Preference Optimization and Proximal Policy Optimization while maintaining computational efficiency.
AINeutralarXiv – CS AI · Jun 46/10
🧠Researchers propose a sparse Mixture-of-Experts (MoE) reward model that learns interpretable, specialized experts for modeling diverse human preferences in RLHF systems. By encouraging sparse routing during training on binary preference data, the approach improves both interpretability and personalization capabilities compared to universal reward function models.
AINeutralarXiv – CS AI · Jun 46/10
🧠Researchers present a new approach to aligning language models with human preferences that works without assuming a specific mathematical relationship between observed preferences and underlying rewards. The method frames policy alignment as a semiparametric optimization problem, enabling more robust policy learning even when the preference model structure is unknown or misspecified.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers introduce Latent Reward Steering (LRS), an inference-time framework that improves reasoning in large language models by optimizing sparse-autoencoder latent states through reward gradients. The method adaptively corrects fragile reasoning states without relying on predefined cognitive behaviors, demonstrating consistent performance improvements across multiple benchmarks.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers propose CARE-RL, a reinforcement learning framework that combines protocol-aware reward generation with capability-aware optimization to address challenges in multi-domain RL systems. The approach achieves improved performance across math, chat, and instruction-following tasks on multiple LLM models, demonstrating advances in making RL more effective across diverse domains.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers propose an improved Nash Learning from Human Feedback (NLHF) algorithm that addresses exploration challenges in preference alignment for large language models. The new method achieves better regret bounds without exponential dependence on regularization parameters and demonstrates empirical improvements when fine-tuning Llama-3-8B.
🧠 Llama
AIBullisharXiv – CS AI · Jun 26/10
🧠Researchers propose Chunk-Level Guided Generation, a training-free method using off-the-shelf large language models to score intermediate reasoning steps during small-model inference for mathematical problem-solving. The approach matches or outperforms specialized reward model-based systems on benchmarks like MATH and GSM8K without requiring expensive step-level training data.
🧠 Llama
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers identify and address Perceptual Judgment Bias in multimodal large language models used as automated evaluators, where these models favor plausible narratives over visually accurate answers when text and images conflict. The team develops a training framework using perceptually perturbed datasets and reward modeling that improves MLLM judges' visual grounding and evaluation consistency across benchmarks.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers introduce Margin-Adaptive Direct Preference Optimization (MADPO), a novel method that improves large language model alignment by using a reward model to apply instance-level adaptive weights to training samples. MADPO addresses limitations in existing approaches like DPO and β-DPO by providing stable, granular control over the learning signal without discarding training data.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers propose Bayesian Non-Negative Reward Model (BNRM), a framework that addresses reward hacking vulnerabilities in reinforcement learning from human feedback (RLHF) systems used to align large language models. The approach combines non-negative factor analysis with preference modeling to create more robust, interpretable reward systems resistant to biases and distribution shifts.
AINeutralarXiv – CS AI · May 296/10
🧠Researchers propose In-Context Reward Adaptation, a transformer-based framework that dynamically models diverse human preferences without costly retraining. By incorporating human response time as an auxiliary signal, the approach enables language models to adapt to unseen preference domains on-the-fly, addressing a critical limitation of static reward models used in RLHF systems.
AIBullisharXiv – CS AI · May 276/10
🧠Researchers propose Tournament-GRPO, a novel reinforcement learning framework that uses group-wise tournament comparisons instead of absolute scoring to improve long-form text generation. By converting rubric-based LLM judgments into relative rewards through competitive rankings, the method achieves 4.52-point improvements over existing approaches on Deep Research Bench benchmarks.
AINeutralarXiv – CS AI · May 126/10
🧠BoostAPR is a new AI framework that improves automated program repair by using dual reward models and reinforcement learning to identify which code edits actually fix bugs. The system achieves significant improvements on multiple benchmarks, including 40.7% on SWE-bench Verified, demonstrating that more granular feedback mechanisms can substantially enhance AI's ability to repair software vulnerabilities.
AINeutralarXiv – CS AI · May 116/10
🧠Researchers propose a method to improve RLHF (Reinforcement Learning from Human Feedback) by treating the rationality parameter as context-dependent rather than fixed, using an LLM-as-judge to detect cognitive biases in human annotations and downweight unreliable comparisons. This approach enables training more robust AI models even when human feedback contains systematic biases.
AIBullisharXiv – CS AI · May 76/10
🧠Researchers have developed methods to efficiently align language models using online natural language feedback in domains where human supervision is limited and difficult to quantify. By iteratively optimizing proxy reward models and collecting fresh expert feedback, the approach recovers 80-100% of full-supervision performance with 3-20x fewer expert samples, demonstrating significant improvements in training data efficiency.
🧠 Haiku
AINeutralarXiv – CS AI · May 46/10
🧠PORTool is a new policy-optimization algorithm that improves how AI agents learn to use external tools by solving the credit-assignment problem in multi-step reasoning tasks. The method uses a rewarded tree structure to assign rewards at individual steps rather than only at outcomes, enabling agents to achieve higher accuracy while reducing unnecessary tool calls.
AINeutralarXiv – CS AI · Apr 206/10
🧠Researchers demonstrate that reward-weighted classifier-free guidance (RCFG) can dynamically adjust autoregressive model outputs to optimize arbitrary reward functions at test time without retraining. Applied to molecular generation, this approach enables real-time optimization of competing objectives and accelerates reinforcement learning convergence when used as a teacher for policy distillation.
AIBullisharXiv – CS AI · Apr 146/10
🧠Researchers propose CPMI, an automated method for training process reward models that reduces annotation costs by 84% and computational overhead by 98% compared to traditional Monte Carlo approaches. The technique uses contrastive mutual information to assign reward scores to reasoning steps in AI chain-of-thought trajectories without expensive human annotation or repeated LLM rollouts.
AIBullisharXiv – CS AI · Mar 166/10
🧠Researchers introduce Visual-ERM, a multimodal reward model that improves vision-to-code tasks by evaluating visual equivalence in rendered outputs rather than relying on text-based rules. The system achieves significant performance gains on chart-to-code tasks (+8.4) and shows consistent improvements across table and SVG parsing applications.
AINeutralarXiv – CS AI · Mar 36/103
🧠Researchers propose rubric-based reward modeling to address reward over-optimization in large language model fine-tuning. The approach focuses on the high-reward tail where models struggle to distinguish excellent responses from merely great ones, using off-policy examples to improve training effectiveness.
AIBullisharXiv – CS AI · Feb 276/107
🧠Researchers propose ContextRL, a new framework that uses context augmentation to improve machine learning model efficiency in knowledge discovery. The framework enables smaller models like Qwen3-VL-8B to achieve performance comparable to much larger 32B models through enhanced reward modeling and multi-turn sampling strategies.
