Tag: faang

This paper was accepted at the Workshop on Unifying Representations in Neural Models (UniReps) at NeurIPS 2025. Activation steering methods in large language models (LLMs) have emerged as an effective way to perform targeted updates to enhance generated language without requiring large amounts of adaptation data. We ask whether the features discovered by activation steering …

Read more “ExpertLens: Activation Steering Features Are Highly Interpretable”

TL;DR: Artificial Intelligence learns patterns from data and uses them to make predictions, generate content, or solve problems. Generative AI, such as ChatGPT or image and video generators, takes this a step further by creating new things, text, art, music, and more, that have never existed before. People often ask: “How does AI actually work?” …

Read more “How does AI work?”

Text Normalization (TN) is a key preprocessing step in Text-to-Speech (TTS) systems, converting written forms into their canonical spoken equivalents. Traditional TN systems can exhibit high accuracy, but involve substantial engineering effort, are difficult to scale, and pose challenges to language coverage, particularly in low-resource settings. We propose PolyNorm, a prompt-based approach to TN using …

Read more “PolyNorm: Few-Shot LLM-Based Text Normalization for Text-to-Speech”

Artificial intelligence is reshaping our world – accelerating discovery, optimising systems, and unlocking new possibilities across every sector. But with its vast potential comes a shared responsibility. AI can be a powerful ally for transforming businesses and reducing cost. It can help organizations minimize carbon emissions, industries manage energy use, and scientists model complex climate …

Read more “Build software sustainably in the AI era”

We introduce Representation Tokenizer (RepTok), a generative modeling framework that represents an image using a single continuous latent token obtained from self-supervised vision transformers. Building on a pre-trained SSL encoder, we fine-tune only the semantic token embedding and pair it with a generative decoder trained jointly using a standard flow matching objective. This adaptation enriches …

Read more “Adapting Self-Supervised Representations as a Latent Space for Efficient Generation”