NVIDIA and Meta Collaborate on Advanced RAG Pipelines with Llama 3.1 and NeMo Retriever NIMs
NVIDIA and Meta introduce scalable agentic RAG pipelines with Llama 3.1 and NeMo Retriever NIMs, optimizing LLM performance and decision-making capabilities.
Enhancing Agent Planning: Insights from LangChain
LangChain explores the limitations and future of planning for agents with LLMs, highlighting cognitive architectures and current fixes.
NVIDIA NeMo Enhances LLM Capabilities with Hybrid State Space Model Integration
NVIDIA NeMo introduces support for hybrid state space models, significantly enhancing the efficiency and capabilities of large language models.
NVIDIA NeMo Curator Enhances Non-English Dataset Preparation for LLM Training
NVIDIA NeMo Curator simplifies the curation of high-quality non-English datasets for LLM training, ensuring better model accuracy and reliability.
WordSmith Enhances Legal AI Operations with LangSmith Integration
WordSmith leverages LangSmith for prototyping, debugging, and evaluating LLM performance, enhancing operations for in-house legal teams.
LangChain: Understanding Cognitive Architecture in AI Systems
Explore the concept of cognitive architecture in AI, outlining various levels of autonomy and their applications in LLM-driven systems.
Understanding the Role and Capabilities of AI Agents
Explore the concept of AI agents, their varying degrees of autonomy, and the importance of agentic behavior in LLM applications, according to LangChain Blog.
Ensuring Integrity: Secure LLM Tokenizers Against Potential Threats
NVIDIA's AI Red Team highlights the risks and mitigation strategies for securing LLM tokenizers to maintain application integrity and prevent exploitation.
LangChain Introduces Self-Improving Evaluators for LLM-as-a-Judge
LangChain's new self-improving evaluators for LLM-as-a-Judge aim to align AI outputs with human preferences, leveraging few-shot learning and user feedback.
IBM Research Unveils Cost-Effective AI Inferencing with Speculative Decoding
IBM Research has developed a speculative decoding technique combined with paged attention to significantly enhance the cost performance of large language model (LLM) inferencing.