Large Language Models (LLMs) excel at tasks like language processing, strategy games, and reasoning but struggle to build generalizable internal representations essential for adaptive decision-making in agents. For agents to effectively navigate complex environments, they must construct reliable world models. While LLMs perform well on specific benchmarks, they often fail to generalize, leading to brittle representations that limit their real-world effectiveness. Understanding how LLMs build internal world models is key to developing agents capable of consistent, adaptive behavior across tasks. We analyze OthelloGPT, a GPT-based model trained on Othello gameplay, as a controlled testbed for studying representation learning. Despite being trained solely on next-token prediction with random valid moves, OthelloGPT shows meaningful layer-wise progression in understanding board state and gameplay. Early layers capture static attributes like board edges, while deeper layers reflect dynamic tile changes. To interpret these representations, we compare Sparse Autoencoders (SAEs) with linear probes, finding that SAEs offer more robust, disentangled insights into compositional features, whereas linear probes mainly detect features useful for classification. We use SAEs to decode features related to tile color and tile stability, a previously unexamined feature that reflects complex gameplay concepts like board control and long-term planning. We study the progression of linear probe accuracy and tile color using both SAE’s and linear probes to compare their effectiveness at capturing what the model is learning. Although we begin with a smaller language model, OthelloGPT, this study establishes a framework for understanding the internal representations learned by GPT models, transformers, and LLMs more broadly. Our code is publicly available: this https URL. — Read More
Daily Archives: January 16, 2025
How should we test AI for human-level intelligence? OpenAI’s o3 electrifies quest
The technology firm OpenAI made headlines last month when its latest experimental chatbot model, o3, achieved a high score on a test that marks progress towards artificial general intelligence (AGI). OpenAI’s o3 scored 87.5%, trouncing the previous best score for an artificial intelligence (AI) system of 55.5%.
This is “a genuine breakthrough”, says AI researcher François Chollet, who created the test, called Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI)1, in 2019 while working at Google, based in Mountain View, California. A high score on the test doesn’t mean that AGI — broadly defined as a computing system that can reason, plan and learn skills as well as humans can — has been achieved, Chollet says, but o3 is “absolutely” capable of reasoning and “has quite substantial generalization power”.
Researchers are bowled over by o3’s performance across a variety of tests, or benchmarks, including the extremely difficult FrontierMath test, announced in November by the virtual research institute Epoch AI. …But many, including Rein, caution that it’s hard to tell whether the ARC-AGI test really measures AI’s capacity to reason and generalize. “ — Read More
Project DIGITS: NVIDIA’s Leap into Personal AI Supercomputing
When you own the platform, you own the experience. That’s why Apple invests so much in the iPhone. That’s what NVIDIA is aiming for with Project DIGITS, unveiled at CES 2025.
Project DIGITS democratizes access to advanced AI computing by introducing a compact and powerful personal AI supercomputer. It’s designed to make it possible for AI researchers, data scientists, students, and even hobbyists to develop, prototype, and fine-tune AI models directly from their desks. While professionals could fine-tune models locally before, they were often constrained by hardware limitations, high costs, or scalability issues. Project DIGITS eliminates these barriers by delivering computing power in a desktop form factor.
As Jensen Huang, founder and CEO of NVIDIA, said in a press release, “AI will be mainstream in every application for every industry. With Project DIGITS, the Grace Blackwell Superchip comes to millions of developers. Placing an AI supercomputer on the desks of every data scientist, AI researcher and student empowers them to engage and shape the age of AI.”
Project DIGITS is also a precursor for how personal computing could fuel the uptake of AI into consumers’ everyday lives in a way that VR devices cannot seem to do – perhaps not today, but sooner than we know. — Read More
DeepSeek-V3
We present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeekMoE architectures, which were thoroughly validated in DeepSeek-V2. Furthermore, DeepSeek-V3 pioneers an auxiliary-loss-free strategy for load balancing and sets a multi-token prediction training objective for stronger performance. We pre-train DeepSeek-V3 on 14.8 trillion diverse and high-quality tokens, followed by Supervised Fine-Tuning and Reinforcement Learning stages to fully harness its capabilities. Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source models and achieves performance comparable to leading closed-source models. Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training. In addition, its training process is remarkably stable. Throughout the entire training process, we did not experience any irrecoverable loss spikes or perform any rollbacks. The model checkpoints are available at this https URL. — Read More
Rick's Cafe AI 