Google DeepMind researchers introduce new benchmark to improve LLM factuality, reduce hallucinations
Based on a new benchmark, Google DeepMind found Gemini 2.0 Flash to be the most factual LLM, with a score ...
Samstag, 28. März 2026
Breaking the data bottleneck: Salesforce’s ProVision speeds multimodal AI training with image scene graphs
Salesforce is using structured representation of image semantics to power programs that synthesize instruction datasets for AI training.Read More
Self-invoking code benchmarks help you decide which LLMs to use for your programming tasks
LLMs are good at coding simple functions. But how good are they at calling their own functions to solve complex ...
LlamaIndex goes beyond RAG so agents can make complex decisions
A new orchestration architecture from LlamaIndex brings more context and decision-making capabilities to AI agents.Read More
Microsoft’s new rStar-Math technique upgrades small models to outperform OpenAI’s o1-preview at math problems
Phi-4 and an rStar-Math paper suggest that compact, specialized models can provide powerful alternatives to the industry’s largest systems.Read More
Diffbot’s AI model doesn’t guess—it knows, thanks to a trillion-fact knowledge graph
Diffbot challenges AI giants with a breakthrough open-source language model that outperforms ChatGPT and Google Gemini by prioritizing real-time fact ...
How Narvar is using AI and data to enhance post-purchase customer experiences
How Narvar's new tech is using AI to harness the power of 42 billion consumer interactions a year to improve ...
Cohere just launched ‘North’, its biggest AI bet yet for privacy-focused enterprises
Cohere launches North, a secure enterprise AI platform outperforming Microsoft Copilot and Google Vertex AI in testing, with Royal Bank ...
Smoke, reflections and portals: Adobe’s TransPixar takes AI VFX to the next level
Adobe and HKUST researchers unveil TransPixar, a breakthrough AI system that creates transparent visual effects from text prompts.Read More
Microsoft makes powerful Phi-4 model fully open source on Hugging Face
Phi-4 demonstrates that smaller, well-designed models can achieve comparable or superior results to larger models.Read More
