Related papers: Behavioral Canaries: Auditing Private Retrieved Co…

While large language models (LLMs) have demonstrated strong performance on complex reasoning tasks such as competitive programming (CP), existing methods predominantly focus on single-attempt settings, overlooking their capacity for…

Generalizable alignment is a core challenge for deploying Large Language Models (LLMs) safely in real-world NLP applications. Current alignment methods, including Reinforcement Learning from Human Feedback (RLHF), often fail to guarantee…

Large language models (LLMs) acquire general linguistic knowledge from massive-scale pretraining. However, pretraining data mainly comprised of web-crawled texts contain undesirable social biases which can be perpetuated or even amplified…

Evaluations of unlearning on reasoning models sometimes show a bypass pattern. The answer side looks unlearned, but the model's own thinking trace keeps emitting the forgotten content, and the gap is taken as evidence that the weights still…

Reinforcement learning (RL) plays a central role in improving the reasoning and alignment of large language models, yet its efficiency critically depends on how training data are selected. Existing online selection strategies predominantly…

Recent strides in large language models (LLMs) have yielded remarkable performance, leveraging reinforcement learning from human feedback (RLHF) to significantly enhance generation and alignment capabilities. However, RLHF encounters…

Large Language Models (LLMs) as clinical agents require careful behavioral adaptation. While adept at reactive tasks (e.g., diagnosis reasoning), LLMs often struggle with proactive engagement, like unprompted identification of critical…

Attribution theory explains how individuals interpret and attribute others' behavior in a social context by employing personal (dispositional) and impersonal (situational) causality. Large Language Models (LLMs), trained on human-generated…

Large language models (LLMs) can be controlled at inference time through prompts (in-context learning) and internal activations (activation steering). Different accounts have been proposed to explain these methods, yet their common goal of…

Diffusion models are powerful generative models that allow for precise control over the characteristics of the generated samples. While these diffusion models trained on large datasets have achieved success, there is often a need to…

Reinforcement learning (RL) algorithms have been successfully used to develop control policies for dynamical systems. For many such systems, these policies are trained in a simulated environment. Due to discrepancies between the simulated…

Mitigating hallucinations in Large Language Models (LLMs) is critical for their reliable deployment. Existing methods typically fine-tune LLMs to abstain from answering questions beyond their knowledge scope. However, these methods often…

Instruction tuning increasingly relies on LLM-based prompt refinement, where prompts in the training corpus are selectively rewritten by an external refiner to improve clarity and instruction alignment. This motivates an instance-level…

We propose RLAnything, a reinforcement learning framework that dynamically forges environment, policy, and reward models through closed-loop optimization, amplifying learning signals and strengthening the overall RL system for any LLM or…

Large Language Models (LLMs) are increasingly exposed to adaptive jailbreaking, particularly in high-stakes Chemical, Biological, Radiological, and Nuclear (CBRN) domains. Although streaming probes enable real-time monitoring, they still…

Despite the promise of RLHF in aligning LLMs with human preferences, it often leads to superficial alignment, prioritizing stylistic changes over improving downstream performance of LLMs. Underspecified preferences could obscure directions…

Reasoning models leveraging long chains of thought employ various cognitive skills, such as verification of their answers, backtracking, retrying by an alternate method, and more. Previous work has shown that when a base language model…

Reinforcement fine-tuning (RFT) has shown promise for training LLM agents to perform multi-turn decision-making based on environment feedback. However, most existing evaluations remain largely in-domain: training and testing are conducted…

While Large Language Models (LLMs) exhibit remarkable capabilities, they also introduce significant safety and privacy risks. Current mitigation strategies often fail to preserve contextual reasoning capabilities in risky scenarios.…

Recent research has turned to Reinforcement Learning (RL) to solve challenging decision problems, as an alternative to hand-tuned heuristics. RL can learn good policies without the need for modeling the environment's dynamics. Despite this…