相关论文: SCOPE: Intrinsic Semantic Space Control for Mitiga…

Sparse autoencoders (SAEs) extract millions of interpretable features from a language model, but flat feature inventories aren't very useful on their own. Domain concepts get mixed with generic and weakly grounded features, while related…

Zero-shot inference, where pre-trained models perform tasks without specific training data, is an exciting emergent ability of large models like CLIP. Although there has been considerable exploration into enhancing zero-shot abilities in…

Steering has emerged as a promising approach in controlling large language models (LLMs) without modifying model parameters. However, most existing steering methods rely on large-scale datasets to learn clear behavioral information, which…

Self-supervised learning (SSL) can be used to solve complex visual tasks without human labels. Self-supervised representations encode useful semantic information about images, and as a result, they have already been used for tasks such as…

The protection of Intellectual Property (IP) for Large Language Models (LLMs) has become a critical concern as model theft and unauthorized commercialization escalate. While adversarial fingerprinting offers a promising black-box solution…

Code authorship attribution is the problem of identifying authors of programming language codes through the stylistic features in their codes, a topic that recently witnessed significant interest with outstanding performance. In this work,…

Deterministically controlling the target generation language of large multilingual language models (LLMs) remains a fundamental challenge, particularly in zero-shot settings where neither explicit language prompts nor fine-tuning are…

Sparse Autoencoders (SAEs) are widely employed for mechanistic interpretability and model steering. Within this context, steering is by design performed by means of decoding altered SAE intermediate representations. This procedure…

The adversarial attacks against deep neural networks on computer vision tasks have spawned many new technologies that help protect models from avoiding false predictions. Recently, word-level adversarial attacks on deep models of Natural…

Symbolic regression (SR) seeks closed-form mathematical expressions that fit observed data. Neural SR methods amortize the search by training an encoder to map observations directly to expressions in a single pass, but this amortized…

While the activations of neurons in deep neural networks usually do not have a simple human-understandable interpretation, sparse autoencoders (SAEs) can be used to transform these activations into a higher-dimensional latent space which…

Typical methods for unsupervised text style transfer often rely on two key ingredients: 1) seeking the explicit disentanglement of the content and the attributes, and 2) troublesome adversarial learning. In this paper, we show that neither…

Sparse autoencoders (SAEs) have been applied to large language models and protein language models, but not systematically to electronic health record (EHR) foundation models. We train TopK SAEs on FlatASCEND, a 14.5-million-parameter…

Recent advances show that semi-supervised implicit representation learning can be achieved through physical constraints like Eikonal equations. However, this scheme has not yet been successfully used for LiDAR point cloud data, due to its…

Sparse Autoencoders (SAEs) have proven to be powerful tools for interpreting neural networks by decomposing hidden representations into disentangled, interpretable features via sparsity constraints. However, conventional SAEs are…

In this work, we introduce Semantic Pyramid AutoEncoder (SPAE) for enabling frozen LLMs to perform both understanding and generation tasks involving non-linguistic modalities such as images or videos. SPAE converts between raw pixels and…

Semi-supervised learning (SSL) has achieved significant progress by leveraging both labeled data and unlabeled data. Existing SSL methods overlook a common real-world scenario when labeled data is extremely scarce, potentially as limited as…

In-context Learning (ICL) is an emerging few-shot learning paradigm on Language Models (LMs) with inner mechanisms un-explored. There are already existing works describing the inner processing of ICL, while they struggle to capture all the…

Large Vision-Language Models (LVLMs) achieve strong performance on many multimodal tasks, but object hallucinations severely undermine their reliability. Most existing studies focus on the text modality, attributing hallucinations to overly…

Multimodal large language models (MLLMs) have revolutionized the landscape of AI, demonstrating impressive capabilities in tackling complex vision and audio-language tasks. However, a critical challenge remains: these models often suffer…