Nonlinear Computation in Deep Linear Networks
OpenAI published a research finding examining how deep linear networks can perform nonlinear computation. The work investigates the theoretical properties of linear neural network architectures and their computational capabilities. This is an older research paper from 2017 that touches on foundational questions about neural network expressivity.
Related guides (1)
Related events (8)
Paper argues Compressed Computation toy model is not computation in superposition
A new arXiv preprint challenges the Compressed Computation (CC) toy model introduced by Braun et al. (2025), which appeared to compute 100 ReLU functions using only 50 neurons. The authors show that apparent performance gains arise from unintended input mixing via a noisy residual stream rather than genuine superposition, with learned neuron directions concentrating in the subspace of the top 50 eigenvalues of the mixing matrix. A semi-non-negative matrix factorization baseline derived purely from the mixing matrix reproduces the qualitative loss profile, supporting the conclusion that CC is not a valid toy model of computation in superposition.
Understanding Neural Networks Through Sparse Circuits
OpenAI has published work on mechanistic interpretability using a sparse model approach aimed at understanding how neural networks reason internally. The research seeks to make AI systems more transparent by identifying sparse circuits within neural networks. This work is positioned as supporting safer and more reliable AI behavior through improved interpretability.
Local linear structures in LLM weights and activations are dynamic, not fixed global directions
A new arXiv paper investigates the nature of linear structures in transformer weights and activations, finding strong local low-rank task-gradient structure but rejecting the hypothesis that fixed task planes exist. The authors show that useful bases drift substantially within 100 optimization steps, yet early recovery updates form a trajectory-prefix basis capturing 77% of LoRA recovery displacement. They also establish a formal connection between parameter perturbations and activation steering, finding a 0.58 cosine similarity between gradient-step-induced activation shifts and CAA steering vectors, suggesting linear structures are evolving local geometries rather than stable global task directions.
AI and Compute: OpenAI Analysis of Exponential Growth in Training Compute Since 2012
OpenAI published an analysis in May 2018 showing that compute used in the largest AI training runs has been doubling every 3.4 months since 2012, far outpacing Moore's Law's 2-year doubling period. Over the 2012–2018 period, this metric grew by more than 300,000x. The analysis frames compute scaling as a key driver of AI progress and argues for preparing for systems with capabilities well beyond those of the time.
Measuring Open-Source Llama Nemotron Models on DeepResearch Bench
NVIDIA evaluates its open-source Llama Nemotron models on the DeepResearch Bench, a benchmark designed to assess deep research agent capabilities. The post appears to report competitive performance of the Nemotron models in agentic research tasks. This is relevant to the ongoing development of open-weights models capable of multi-step research and reasoning workflows.
Scaling Laws for Neural Language Models
OpenAI published foundational research establishing empirical scaling laws for neural language models, showing that model performance scales predictably with compute, data, and parameters. The work demonstrated power-law relationships between these factors and loss, providing a principled framework for allocating training resources. This paper became a cornerstone of modern large language model development strategy.
Computational limitations in robust classification and win-win results
OpenAI published research examining computational limitations in robust classification, exploring theoretical bounds on adversarially robust machine learning. The work investigates so-called 'win-win' results where both standard and robust accuracy can be achieved simultaneously. This is a foundational safety and robustness research contribution from 2019, addressing hardness results in adversarial ML.
Dynamic short convolutions yield 1.33–1.60× compute advantage over standard Transformers
A new arXiv preprint introduces dynamic short convolutions as an architectural primitive for Transformers, using input-dependent filters to combine locality bias with increased expressivity. Experiments across 150M–2B parameter language models show consistent perplexity improvements over standard Transformers and static convolution variants, with scaling-law fits indicating a 1.33× compute advantage when applied to key/query/value vectors and 1.60× when added after every linear layer. The technique also improves linear RNNs (Mamba-2, Gated DeltaNet) and mixture-of-experts architectures, with custom Triton kernels making training practical.