相关论文: Probing RLVR training instability through the lens…

Reinforcement learning from human feedback (RLHF) has emerged as a powerful technique to make large language models (LLMs) more capable in complex settings. RLHF proceeds as collecting human preference data, training a reward model on said…

Reinforcement Learning with Verifiable Rewards (RLVR) is widely used to improve reasoning in multiple domains, yet outcome-only scalar rewards are often sparse and uninformative, especially on failed samples, where they merely indicate…

Hyperparameter sensitivity in Deep Reinforcement Learning (RL) is often accepted as unavoidable. However, it remains unclear whether it is intrinsic to the RL problem or exacerbated by specific training mechanisms. We investigate this…

Reinforcement Learning with Verifiable Rewards (RLVR) is an effective paradigm for improving the reasoning capabilities of large language models. However, existing RLVR methods utilize rollouts in an indiscriminate and short-horizon manner:…

Robustness to noise is of utmost importance in reinforcement learning systems, particularly in military contexts where high stakes and uncertain environments prevail. Noise and uncertainty are inherent features of military operations,…

Reinforcement learning with verifiable reward (RLVR) has become a promising paradigm for post-training large language models (LLMs) to improve their reasoning capability. However, when the rollout accuracy is low on hard problems, the…

In this work, we study the issue of reward hacking on the response length, a challenge emerging in Reinforcement Learning from Human Feedback (RLHF) on LLMs. A well-formatted, verbose but less helpful response from the LLMs can often…

Modulation instability is a phenomenon of spontaneous pattern formation in nonlinear media, oftentimes leading to an unpredictable behaviour and a degradation of a signal of interest. We propose an approach based on reinforcement learning…

Recent studies have shown that reinforcement learning (RL) models are vulnerable in various noisy scenarios. For instance, the observed reward channel is often subject to noise in practice (e.g., when rewards are collected through sensors),…

Reinforcement Learning from Verifiable Rewards (RLVR) has emerged as a powerful paradigm for enhancing Large Language Models (LLMs), exemplified by the success of OpenAI's o-series. In RLVR, rewards are derived from verifiable signals-such…

We study a class of reinforcement learning problems where the reward signals for policy learning are generated by an internal reward model that is dependent on and jointly optimized with the policy. This interdependence between the policy…

Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective for Large Language Model (LLM) reasoning, yet current methods face key challenges in resource allocation and policy optimization dynamics: (i) uniform rollout…

Aligning large language models to human preferences is inherently multidimensional, yet most pipelines collapse heterogeneous signals into a single optimizeable objective. We seek to answer what it would take to simultaneously align a model…

When learning behavior, training data is often generated by the learner itself; this can result in unstable training dynamics, and this problem has particularly important applications in safety-sensitive real-world control tasks such as…

Reinforcement learning with verifiable rewards (RLVR) has demonstrated significant success in enhancing mathematical reasoning and coding performance of large language models (LLMs), especially when structured reference answers are…

Reinforcement fine-tuning with verifiable rewards (RLVR) has emerged as a powerful paradigm for equipping large vision-language models (LVLMs) with agentic capabilities such as tool use and multi-step reasoning. Despite striking empirical…

Large Language Models (LLMs) have recently improved mathematical reasoning through Reinforcement Learning with Verifiable Reward (RLVR). However, existing RLVR algorithms require large query budgets, making annotation costly. We investigate…

Reward hacking arises when a model improves a proxy reward by exploiting shortcuts rather than solving the intended task. We study this failure mode through the geometry of reinforcement learning updates in language models and argue that…

Mode-dependent architectural components (layers that behave differently during training and evaluation, such as Batch Normalization or dropout) are commonly used in visual reinforcement learning but can destabilize on-policy optimization.…

Recently, reinforcement learning with verifiable rewards (RLVR) has been widely used for enhancing the reasoning abilities of large language models (LLMs). A core challenge in RLVR involves managing the exchange between entropy and…