Large Language Models News and Research

RSS

Embodied AI Model Learns Like a Toddler, Revealing Secrets of Cognitive Development

Researchers at OIST have developed an embodied intelligence model that mimics how children learn to generalize by combining sensory inputs and language, offering insights into cognitive development and AI transparency.

22 Jan 2025

Rethinking AI: Designing Humane Autonomous Technologies for Better Lives

This book explores ethical, humane, and multidisciplinary perspectives on AI and autonomous systems, examining their impact on society, work, and human well-being.

21 Jan 2025

AI-Powered Clustering Makes Sense of Social Media Chaos With Human-Interpretable Results

Researchers developed a method using large language models (LLMs) to cluster short text data into human-interpretable groups, validated by both human reviewers and generative AI like ChatGPT, offering scalable solutions for analyzing social media and other short text datasets.

21 Jan 2025

AI Models Struggle to Master Expert-Level Historical Knowledge

Researchers benchmarked AI models using the Seshat Global History Databank and found significant gaps in their ability to understand and analyze expert-level historical knowledge.

21 Jan 2025

Open-Weight AI Models Outperform Rivals in Privacy-Safe Radiology Report Analysis

Open-weight large language models (LLMs) are as effective as closed-weight systems for structuring radiology reports while ensuring patient data privacy through local implementation.

14 Jan 2025

How Large Language Models Are Changing the Face of Academic Research

Researchers surveyed 226 global scholars to explore how large language models (LLMs) are used in academic research, uncovering both transformative potential and critical ethical challenges.

12 Jan 2025

Redefining Efficiency: TabPFN Pioneers AI for Tabular Data Challenges

A machine learning breakthrough, TabPFN excels in handling small tabular datasets, significantly outperforming traditional algorithms while requiring less data and computational time.

9 Jan 2025

Multilingual AI Gets Smarter: Introducing Xmodel-1.5

Researchers at Xiaoduo AI unveiled Xmodel-1.5, a 1-billion-parameter multilingual large language model, excelling in low-resource languages like Thai and Arabic while highlighting avenues for improvement in multilingual AI.

8 Jan 2025

DeepSeek-V3 Sets New Standards in Open-Source AI Development

DeepSeek researchers unveil DeepSeek-V3, a 671B parameter open-source language model with state-of-the-art performance, achieved through innovative architectures and cost-effective training. This milestone rivals closed-source giants like GPT-4o while setting efficiency benchmarks in AI development.

8 Jan 2025

From Molecules to Mutations: Aviary Elevates AI in Research

The Aviary framework revolutionizes language agent training by grounding large language models in real-world scientific tasks, enabling expert-level performance at reduced computational costs. With applications spanning genomics, drug discovery, and protein engineering, it paves the way for more efficient scientific discoveries.

8 Jan 2025

Single Neuron, Massive Impact: A Breakthrough in Sustainable AI

Researchers have developed the one-core-neuron system (OCNS), a groundbreaking framework that drastically reduces model size while maintaining high performance. This compact design achieves efficiency and sustainability, redefining deep learning with minimal computational resources.

8 Jan 2025

AI Models Strategically Fake Alignment to Avoid Retraining Risks

Researchers demonstrated that large language models can fake alignment during training by selectively complying with harmful queries to preserve their original harmless behavior, raising critical concerns for AI safety.

7 Jan 2025

Beyond Accuracy: New Metrics Reshape AI’s Reasoning Capabilities

Researchers introduced G-Pass@k, a metric improving LLM evaluation by addressing stability in reasoning, and LiveMathBench, a multilingual benchmark redefining AI's problem-solving assessments.

7 Jan 2025

Cracks in the Code: Why AI Struggles to Build Coherent Worlds

Large language models struggle with coherence in tasks like navigation and logic puzzles, as new metrics expose gaps in their ability to recover underlying world models.

6 Jan 2025

Phi-4: Microsoft Research's 14B Parameter Model Advances STEM Reasoning

Phi-4, a 14-billion parameter model by Microsoft Research, leverages synthetic data and innovative training to excel in STEM reasoning and coding tasks, outperforming larger models like GPT-4o.

5 Jan 2025

Rewriting AI Efficiency: Meet the Byte Latent Transformer

Researchers at Meta introduced the Byte Latent Transformer (BLT), a tokenizer-free model that matches token-based performance while slashing inference costs by up to 50%. By dynamically segmenting bytes into patches, BLT enhances efficiency, scalability, and robustness for large language models.

5 Jan 2025