每天自动聚合 AI 领域最新动态
Computer-use agents increasingly tackle long-horizon tasks, where running a frontier model at every step is expensive...
Sparse autoencoders (SAEs) are typically trained on hundreds of millions to billions of activation tokens, yet little...
Sparse autoencoders (SAEs) represent model activations with sparse, linearly decoded features, but the downstream eff...
Modern generative models produce high-quality individual outputs but struggle to sample diversely, even when explicit...
Time Series Language Models (TSLMs) enable language-based understanding and reasoning over temporal data, yet a funda...
Image editing models produce impressive results for individual edits, but their quality often deteriorates over multi...
Continuous diffusion language models (CDLMs) offer an alternative approach to parallel language generation; however, ...
Deep search agents solve knowledge-intensive tasks through long-horizon, multi-turn interactions, where assigning cre...
Large language models (LLMs) can often tell when they are being tested. This ability, known as evaluation awareness, ...
How can robots autonomously collect the contact-rich experience needed to learn dexterous manipulation? We present DS...
We introduce 3DCodeVerse, a research ecosystem that enables coding agents to create editable and verifiable 3D conten...
Recent 3D world models increasingly use features from geometry foundation models (GFMs) as predictive world states, b...