Search papers, labs, and topics across Lattice.
6
1
11
10
Atria Dawn Preview is introduced, a foundation agentic language model designed for scientific research and engineering workflows, with the goal of expanding the frontier of agent productivity in the real world.
Recycling just four transformer layers through twelve recurrent iterations matches full-depth baselines on core linguistic benchmarks, exposing precisely where compute can—and cannot—substitute for raw parameter capacity.
Natural-language critiques can transform how we evaluate and optimize song generation models, leading to more human-aligned outputs.
Today's best language models can barely make sense of your messy group chats and fragmented digital life, achieving only 19% accuracy on a new benchmark of real-world reasoning.
Retrofit your VLMs with Multi-Head Latent Attention (MLA) for faster inference and smaller memory footprint, without costly pretraining, using this parameter-efficient conversion framework.
Finally, a fully open-source, reproducible system for long-form song generation is here, complete with licensed data, code, and a Qwen-based model that rivals closed-source systems.
