Large language models can answer scientific questions, yet a correct output does not reveal whether the model represents or uses the governing physics. Here we show…
We present AutoIndex, a framework for learning representation programs: executable transformations that map raw documents into the representations exposed to a retrieval system. Rather than tuning…
Diffusion Language Models (DLMs) have recently emerged as a promising alternative to autoregressive models. Unlike standard diffusion-based approaches, DLMs are not explicitly conditioned on a timestep,…
Modern ASR models trained on heterogeneously annotated data treat transcription style (verbatim vs. intended) as an uncontrolled latent variable, causing measurable decoding instability, evaluation confounding (up…
Multimodal humor in memes, cartoons, and comics remains difficult for AI systems because intended meaning depends on non-literal mechanisms, shared cultural knowledge, and communicative intent rather…
Accurate agricultural field boundary delineation at large scale is a foundational task for food security, supply chain transparency, and carbon accounting. While vision foundation models like…
Optimizer state is the largest single line item in the memory budget of mixture-of-experts (MoE) training: on a 6.78B-parameter MoE language model, AdamW keeps 50.6 GB…
Teaching videos are becoming a major medium for education, creating a growing need for scalable evaluation of their pedagogical quality. Existing automatic judges do not fully…
Generative world renderer AlayaRenderer receives structured world states exported from physics engines and synthesizes RGB frames. Unlike models that generate frames from text/control-hints prompts, AlayaRenderer preserves…
LLM agent failures are difficult to debug because the step where an error surfaces is often not the one that caused it. Existing observability tools replay…