Related papers: Safe RLHF-V: Safe Reinforcement Learning from Mult…

The safety mechanisms of large language models (LLMs) exhibit notable fragility, as even fine-tuning on datasets without harmful content may still undermine their safety capabilities. Meanwhile, existing safety alignment methods…

Reinforcement learning from human feedback (RLHF) has proven effective in enhancing the instruction-following capabilities of large language models; however, it remains underexplored in the cross-modality domain. As the number of modalities…

Recent advancements in open-source multi-modal large language models (MLLMs) have primarily focused on enhancing foundational capabilities, leaving a significant gap in human preference alignment. This paper introduces OmniAlign-V, a…

Reinforcement Learning from Human Feedback (RLHF) aligns language models to human preferences by employing a singular reward model derived from preference data. However, such an approach overlooks the rich diversity of human preferences…

Achieving robust safety alignment in large language models (LLMs) while preserving their utility remains a fundamental challenge. Existing approaches often struggle to balance comprehensive safety with fine-grained controllability at the…

The emergence of Vision Language Models (VLMs) has brought unprecedented advances in understanding multimodal information. The combination of textual and visual semantics in VLMs is highly complex and diverse, making the safety alignment of…

The deployment of large language models (LLMs) raises significant ethical and safety concerns. While LLM alignment techniques are adopted to improve model safety and trustworthiness, adversaries can exploit these techniques to undermine…

Guardrails are critical for the safe deployment of Large Language Models (LLMs)-powered software. Unlike traditional rule-based systems with limited, predefined input-output spaces that inherently constrain unsafe behavior, LLMs enable…

Reinforcement learning from human feedback (RLHF) is a key paradigm for aligning large language models (LLMs) with human values, yet the reward models at its core remain largely opaque. In this work, we present Sparse Autoencoder For…

We apply preference modeling and reinforcement learning from human feedback (RLHF) to finetune language models to act as helpful and harmless assistants. We find this alignment training improves performance on almost all NLP evaluations,…

This paper presents a systematic evaluation of Large Language Models' (LLMs) behavior on long-tail distributed (encrypted) texts and their safety implications. We introduce a two-dimensional framework for assessing LLM safety: (1)…

Large Language Models (LLMs) pose a significant risk of safety misalignment after finetuning, as models can be compromised by both explicitly and implicitly harmful data. Even some seemingly benign data can inadvertently steer a model…

Large language models (LLMs) are increasingly deployed behind safety guardrails such as system prompts and content filters, especially in settings where product teams cannot modify model weights. In practice these guardrails are typically…

Safeguarding vision-language models (VLMs) is a critical challenge, as existing methods often suffer from over-defense, which harms utility, or rely on shallow alignment, failing to detect complex threats that require deep reasoning. To…

Reinforcement Learning (RL) has emerged as a transformative approach for aligning and enhancing Large Language Models (LLMs), addressing critical challenges in instruction following, ethical alignment, and reasoning capabilities. This…

Omni-modal Large Language Models (OLLMs) that process text, images, videos, and audio introduce new challenges for safety and value guardrails in human-AI interaction. Prior guardrail research largely targets unimodal settings and typically…

We study estimation and statistical inference for reward models used in aligning large language models (LLMs). A key component of LLM alignment is reinforcement learning from human feedback (RLHF), where humans compare pairs of…

The rapid advancement of Large Vision-Language Models (LVLMs) is increasingly accompanied by unauthorized scraping and training on multimodal web data, posing severe copyright and privacy risks to data owners. Existing countermeasures, such…

Given the growing influence of language model-based agents on high-stakes societal decisions, from public policy to healthcare, ensuring their beneficial impact requires understanding the far-reaching implications of their suggestions. We…

Recent advancements in LLMs have shown their significant potential in tasks like text summarization and generation. Yet, they often encounter difficulty while solving complex physics problems that require arithmetic calculation and a good…