15 accepted papers
As frontier AI systems become increasingly capable, concerns about deceptive behaviors have intensified. Unlike hallucinations, which stem from capability limitations, deception involves strategically misleading responses despite correct internal representations. While prior work has primarily studi…
Recent advances in generative audio models have enabled high-fidelity environmental sound synthesis, raising serious concerns for audio security. The ESDD 2026 Challenge therefore addresses environmental sound deepfake detection under unseen generators (Track 1) and black-box low-resource detection…
As the size of large language models grows exponentially, GPU memory has become a bottleneck for adapting these models to downstream tasks. In this paper, we aim to push the limits of memory-efficient training by minimizing memory usage on model weights, gradients, and optimizer states, within a uni…
As LLM-based agents increasingly operate in high-stakes domains with real-world consequences, ensuring their behavioral safety becomes paramount. The dominant oversight paradigm, *LLM-as-a-Judge*, faces a fundamental dilemma: how can probabilistic systems reliably supervise other probabilistic syste…
Although Multimodal Large Language Models (MLLMs) have demonstrated proficiency in video captioning, practical applications require captions that follow specific user instructions rather than generating exhaustive, unconstrained descriptions. Current benchmarks, however, primarily assess descriptiv…
Large Audio Language Models (LALMs) represent an important frontier in multimodal AI, addressing diverse audio tasks. Recently, post-training of LALMs has received increasing attention due to significant performance improvements over foundation models. While single-stage post-training such as reinfo…
Training multi-modal large language models (MLLMs) that align with human intentions is a long-term challenge. Traditional score-only reward models for alignment suffer from low accuracy, weak generalization, and poor interpretability, blocking the progress of alignment methods, \textit{e.g.,} reinfo…
As multimodal large models (MLLMs) continue to advance across challenging tasks, a key question emerges: \textbf{\textit{What essential capabilities are still missing? }} A critical aspect of human learning is continuous interaction with the environment -- not limited to language, but also involving…
Large language models (LLMs) may exhibit unintended or undesirable behaviors. Recent works have concentrated on aligning LLMs to mitigate harmful outputs. Despite these efforts, some anomalies indicate that even a well-conducted alignment process can be easily circumvented, whether intentionally or…
In this work, we introduce the PKU-SafeRLHF dataset, designed to promote research on safety alignment in large language models (LLMs). As a sibling project to SafeRLHF and BeaverTails, we separate annotations of helpfulness and harmlessness for question-answering pairs, providing distinct perspectiv…
Existing alignment methods share a common topology of information flow, where reward information is collected from humans, modeled with preference learning, and used to tune language models. However, this shared topology has not been systematically characterized, nor have its alternatives been thoro…
Multimodal large language models (MLLMs) are essential for building general-purpose AI assistants; however, they pose increasing safety risks. How can we ensure safety alignment of MLLMs to prevent undesired behaviors? Going further, it is critical to explore how to fine-tune MLLMs to preserve capab…
Aligning the behavior of Large language models (LLMs) with human intentions and values remains a critical challenge. Reinforcement learning from human feedback (RLHF) aligns LLMs by training a reward model (RM) on human preferences and fine-tuning the LLMs to maximize RM feedback. Despite its effect…
Artificial intelligence (AI) systems possess significant potential to drive societal progress. However, their deployment often faces obstacles due to substantial safety concerns. Safe reinforcement learning (SafeRL) emerges as a solution to optimize policies while simultaneously adhering to multiple…
In a real-world setting, machine needs to continuously recognize new categories without forgetting. However, the number of new categories may be small. For some difficult categories, even humans cannot recognize only based on few-shot examples. To address the above issues, an innovative uncertainty-…