Arnau Marin-Llobet

Arnau Marin-Llobet, Simon Henniger, Mahzarin R. Banaji

Alignment teaches vision-language models (VLMs) to avoid expressing demographic biases, and when gender is clearly visible they largely succeed. Far less is known about ambiguous inputs (a worker in full gear, a figure seen from behind) cases common in practice yet rarely studied. We find that minimal prompting pressure exposes occupation-gender defaults when prompting ambiguous input images, with models collapsing to male even for strongly female-stereotyped occupations. But do these outputs reflect what models actually encode internally? We introduce LALS (Latent Association Leaning Score), a zero-shot metric that projects visual-token activations into the model's text-embedding space to measure concept associations per token and layer. Across 15 occupations, over 800 gender-ambiguous images, and four VLMs, internal representations and outputs are systematically decoupled: models often encode a female association internally yet output male. Layer-wise analysis reveals an asymmetric filter -- male signal amplifies end-to-end while female signal peaks mid-network and is suppressed before generation -- and a color ablation shows that culturally loaded visual cues such as clothing color further modulate these internal associations.

Arnau Marin-Llobet, Javier Ferrando

We introduce an autonomous multiagent framework for mechanistic interpretability that automates both explaining and finding internal features in large language models. The system runs two coupled loops: (1) explanation refinement, where an agent proposes competing hypotheses and iteratively tests them with targeted prompt controls and a multi-metric evaluation; and (2) feature discovery, where an agent generates prompt sets, constructs a k-nearest-neighbor graph in activation space, and retrieves candidate features using statistical separability and semantic coherence criteria. On Gemma-2 family models and MLP neurons in weight-sparse transformers, our agent improves over one-shot auto-interpretations, discovers language-specific and safety-relevant features, and produces auditable explanation traces, showing that agent-driven empirical loops yield sharper and more falsifiable explanations than one-shot labels.

Arnau Marin-Llobet, Arnau Manasanch, Maria V. Sanchez-Vives

The study of brain states, ranging from highly synchronous to asynchronous neuronal patterns like the sleep-wake cycle, is fundamental for assessing the brain's spatiotemporal dynamics and their close connection to behavior. However, the development of new techniques to accurately identify them still remains a challenge, as these are often compromised by the presence of noise, artifacts, and suboptimal recording quality. In this study, we propose a two-stage computational framework combining Hopfield Networks for artifact data preprocessing with Convolutional Neural Networks (CNNs) for classification of brain states in rat neural recordings under different levels of anesthesia. To evaluate the robustness of our framework, we deliberately introduced noise artifacts into the neural recordings. We evaluated our hybrid Hopfield-CNN pipeline by benchmarking it against two comparative models: a standalone CNN handling the same noisy inputs, and another CNN trained and tested on artifact-free data. Performance across various levels of data compression and noise intensities showed that our framework can effectively mitigate artifacts, allowing the model to reach parity with the clean-data CNN at lower noise levels. Although this study mainly benefits small-scale experiments, the findings highlight the necessity for advanced deep learning and Hopfield Network models to improve scalability and robustness in diverse real-world settings.

Xinhe Zhang, Yuyang Zhang, Pengfei Jin et al.

High-resolution 3D medical image generation remains challenging because fully volumetric models are computationally expensive, while efficient 2D slice generators often fail to preserve anatomical consistency across the third dimension. We propose LiFT, a framework for Lifted inter-slice Feature Trajectories that factorizes 3D volume synthesis into per-slice image generation and inter-slice trajectory learning. Rather than modeling the volumetric distribution end-to-end, LiFT treats a volume as an ordered trajectory in feature space, capturing how anatomical structures appear, transform, and disappear across depth. A tri-planar drifting loss aligns the trajectory of generated slices with the trajectories of real volumes, enabling distributional learning over inter-slice progressions in unconditional generation; in paired translation, a bidirectional $z$-context mixer trained against the registered target supplies through-plane coherence while preserving per-slice fidelity. We evaluate LiFT on BraTS 2023 (unconditional and missing-modality MR) and SynthRAD2023 (MR-to-CT). Across these settings, LiFT preserves per-slice quality, approaches the reported cWDM missing-MR reconstruction quality at $\sim$$135\times$ lower inference cost (without formal equivalence testing), and improves through-plane coherence on MR-to-CT relative to a no-mapper ablation, demonstrating that lightweight inter-slice trajectory learning is a viable route to high-resolution 3D medical synthesis.

Arnau Marin-Llobet, Arnau Manasanch, Sergio Sanchez-Manso et al.

This study investigates the application of Riemannian geometry-based methods for brain decoding using invasive electrophysiological recordings. Although previously employed in non-invasive, the utility of Riemannian geometry for invasive datasets, which are typically smaller and scarcer, remains less explored. Here, we propose a Minimum Distance to Mean (MDM) classifier using a Riemannian geometry approach based on covariance matrices extracted from intracortical Local Field Potential (LFP) recordings across various regions during different brain state dynamics. For benchmarking, we evaluated the performance of our approach against Convolutional Neural Networks (CNNs) and Euclidean MDM classifiers. Our results indicate that the Riemannian geometry-based classification not only achieves a superior mean F1 macro-averaged score across different channel configurations but also requires up to two orders of magnitude less computational training time. Additionally, the geometric framework reveals distinct spatial contributions of brain regions across varying brain states, suggesting a state-dependent organization that traditional time series-based methods often fail to capture. Our findings align with previous studies supporting the efficacy of geometry-based methods and extending their application to invasive brain recordings, highlighting their potential for broader clinical use, such as brain computer interface applications.