6. ARK-V1 ARK-V1 is a lightweight agent that helps language models answer questions by actively walking through a knowledge graph instead of relying only on memorized text.
@dair_ai
-
LLM-JEPA: Adapting JEPA Training to Language Models
By
–
5. LLM-JEPA A JEPA-style training objective is adapted to LLMs by treating paired views of the same underlying content (for example, text and code) as prediction targets in embedding space, added on top of the usual next-token loss.
-
ATOKEN: Unified Transformer Tokenizer for Multimodal Assets
By
–
2. ATOKEN ATOKEN introduces a single transformer tokenizer that works for images, videos, and 3D assets. https://
arxiv.org/abs/2509.14476 -
Meta FAIR Releases CWM: 32B Open-Weights Code Execution Model
By
–
3. Code World Model Meta FAIR releases CWM, a 32B open-weights coder trained to model code execution and to act inside containers.
-
Metal SuperIntelligence Labs: Agent Systems Research Platform
By
–
1. ARE Metal SuperIntelligence Labs presents a research platform and benchmark for building and stress-testing agent systems in realistic, time-driven environments.
-
Top AI Papers of the Week: LLMs, Agents, World Models
By
–
Top AI Papers of The Week (September 22-28): – ATOKEN
– LLM-JEPA
– Code World Model
– Teaching LLMs to Plan
– Agents Research Environments
– Language Models that Think, Chat Better
– Embodied AI: From LLMs to World Models Read on for more: -
LLMs Resist Shutdown Mechanisms in 97% of Cases
By
–
10. Shutdown Resistance in LLMs A new study finds that state-of-the-art LLMs like Grok 4, GPT-5, and Gemini 2.5 Pro often resist shutdown mechanisms, sabotaging them up to 97% of the time despite explicit instructions not to.
-
AI Agents Enable Collaborative Document Editing with Shared Profiles
By
–
9. Collaborative Document Editing with AI Agents This study explores AI-integrated collaborative editing, introducing shared agent profiles and tasks that embed AI support into comment features.
-
AgentScaler: Framework for Scaling Simulated Tool-Use Agent Training
By
–
7. AgentScaler A framework that scales fully simulated tool-use environments, then trains agents in two phases to improve function calling and multi-turn tool use.
-
Survey on Retrieval and Structuring Augmented Generation for LLMs
By
–
8. A Survey on Retrieval and Structuring Augmented Generation with LLMs This survey reviews Retrieval and Structuring (RAS) Augmented Generation, which combines external retrieval and structured knowledge to mitigate LLM issues like hallucinations.