Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
3
3
5
5
The DeepSeek-V4.1-Flash model, a multimodal Mixture-of-Experts model with 552B backbone parameters and support for contexts of up to one million tokens, is introduced, substantially improving cost efficiency for agentic workloads and pushing the limits of KV cache compression.
Quantizing user preferences into discrete tokens unlocks personalized multimodal content generation with improved consistency between modalities.
Forget task-specific fine-tuning: TSEmbed unlocks SOTA multimodal embeddings by disentangling task objectives with a Mixture-of-Experts and a novel expert-aware negative sampling strategy.