SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning

Rethinking Shrinkage Bias in LLM FP4 Pretraining: Geometric Origin, Systemic Impact, and UFP4 Recipe

Hi there 👋 Welcome to inclusionAI, home for Ant Group's Artificial General Intelligence (AGI) initiative. In here you can find Large Language Models (LLM), Reinforcement Learning (RL) or other systems related to models.

• Follow our GitHub, Ant Ling team's X Twitter or Ant OSS team's X Twitter for new model releases or open source updates.
• Stay connected at Discord
• Experience the models at ZenMux (recommended) or OpenRouter

🤖 We are also bridging the gap between digital intelligence and the physical world, exploring the frontiers of Embodied AI. Check out Robbyant, an organization under Ant Group dedicated to building the foundational platform for Embodied AI.

👁 https://ospo-insights.oss-cn-hangzhou.aliyuncs.com/hf_iai.gif