Welcome to AGI Advance, Turing’s weekly recap of the most important AI & AGI developments. Each week, we explore frontier AI research, real-world applications, and key industry events shaping the future of artificial intelligence.

Here’s what’s making headlines this week:

What we're thinking

At Turing, we’re constantly pushing the boundaries of AI reasoning and problem-solving. This week, we’re testing how AI can handle unseen knowledge, PhD-level concepts, and spatial reasoning to improve advanced cognition:

Beyond training data: Evaluating AI on the latest research, software updates, and real-world discoveries.
PhD-level challenges: Generating high-complexity questions in biology, chemistry, and physics.
Variants of tough exams: Adapting AI to solve Olympiad-level and entrance exam problems.
3D spatial reasoning: Improving AI’s ability to process complex visual and spatial tasks.

These experiments aim to enhance AI’s problem-solving depth, ensuring models think, reason, and adapt like never before.

What we're reading

We’re diving into three cutting-edge AI research papers this week:

Generalization or Memorization: Data Contamination and Trustworthy Evaluation for Large Language Models
This study introduces CDD, a method for detecting data contamination in large language models, and TED, a technique to improve evaluation reliability.
GPQA: A Graduate-Level Google-Proof Q&A Benchmark
A dataset of 448 high-difficulty, multiple-choice questions in biology, physics, and chemistry, designed to test AI’s expert-level reasoning without relying on web search.
Inherent Limitations of LLMs Regarding Spatial Information
An in-depth look at why GPT-4 and similar models struggle with 2D & 3D route planning, and what it takes to build AI that understands space like humans do.

Each of these papers offers key insights into improving AI’s reliability, reasoning, and real-world effectiveness.

Where we’ll be

Turing will be at two major AI conferences in the coming months—join us to discuss the future of AGI:

AAAI [Philadelphia, USA | Feb 25 – Mar 4]
A leading AI research event covering machine learning, reasoning, robotics, and more.
ICLR 2025 [Singapore | Apr 24 – 28]
A premier conference on deep learning, optimization, and AI theory advancements.

If you’re attending, reach out—we’d love to connect and exchange insights!

Stay ahead with AGI Advance

Turing is leading the charge in bridging AI research with real-world applications. Subscribe to AGI Advance for weekly insights into breakthroughs, research, and industry shifts that matter.

[Subscribe & Read More]