AI BriefingSensenova releases SenseNova-Loop-9B-Preview with 256K context and looped architecture
·2026.10.30 21:34
Key point
The model uses a looped decoder to execute 56 layer passes from 32 distinct layers, supporting text and image inputs for agentic workflows.
Details
- SenseNova-Loop-9B-Preview is a multimodal language model designed for long-context and agentic workflows, accepting text and images to generate text for document understanding, coding, and tool interaction.
- The architecture employs a looped decoder where a contiguous block of 24 layers is executed twice with shared weights, resulting in 56 decoder-layer executions from 32 distinct layers.
- It features two-stream manifold-constrained hyper-connections (mHC) that manage information flow between residual streams and a hybrid attention mechanism interleaving linear and full attention in a four-layer pattern.
- Documented training recipes support maximum sequence lengths of 262,144 tokens, incorporating long sequences, multimodal inputs, and supervised task trajectories.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Researchers demonstrate data-free self-improvement in LLMs using on-policy self-trainingMirendil introduces Fern, its first model with a 10M token context window