Tunneling the Loss Landscape: Bypassing Memorization with Monte Carlo Parameter Swapping

Grokking is a striking phenomenon in neural network training, where a model can undergo a prolonged period of pure memorization before abrupt generalization. While previous works have attempted to interpret it through classical machine learning mechanisms like weight norm, recent research draws an analogy from statistical physics, framing grokking as a form of computational glass relaxation. This theory defines the initial memorization as a result of `fast cooling' where the training loss is reduced so quickly that a glass state is formed, followed by a `slow relaxation' towards final generalization. Although providing a unifying framework for representative grokking theories, this perspective has remained largely at the theoretical on macroscopic level without direct empirical validation on training dynamics. Here we introduce a three-component framework to directly characterize the training dynamics via parameter mobility (PM), and two representative measurements from glassy dynamics: replica correlation (RC) and fractal dimension (FD). We demonstrate that standard optimization presents clear signatures of glass dynamics and inherently traps the grokking network in a kinetic arrested memorization state with a collapsed mobility, strong history dependence, and channel-like motions. This quantitative agreement motivates us to introduce State-Aware Monte Carlo Parameter Swapping (SAM-Swap), an optimization plug-in that can accelerate generalization, inspired by swap Monte Carlo algorithm widely used in glass dynamics. Comparing SAM-Swap, weight decay, and Gaussian gradient noise, we find that accelerated generalization is consistently associated with random exploration in the parameter space, similar to diffusion in physics.

The Grokked Illusion: True Equilibrium Mitigates Catastrophic Forgetting

While neural networks are typically evaluated by their training and test performance, these metrics do not reveal how robust a learned representation is. Recent studies have shown that solutions occupying larger volumes in parameter space, as quantified by Boltzmann entropy, often exhibit superior generalizability compared to those reached by conventional optimization, a phenomenon known as the high entropy advantage. Here we ask whether this advantage persists beyond generalization. Specifically, we investigate models' robustness, the ability to retain the learned knowledge when the model is subsequently trained to acquire new information. Using grokking in modular arithmetic as a controlled setting, we design a noise injection experiment to evaluate the robustness difference between AdamW-trained transformers and high-entropy model sampled from Wang-Landau Molecular Dynamics with identical saturated performance. By forcing both models to fully remember new data with random labels, we find that AdamW-trained models suffer from catastrophic forgetting, with original task test accuracy dropping from 100% to below 75%, whereas the high-entropy models maintain approximately 95% test accuracy. We term this hidden fragility behind apparent generalization the "grokked illusion." Through singular value decomposition of the neural network weights, we discover that high-entropy neural networks possess significantly higher effective rank in attention and MLP layers both before and after noise injection, indicating richer feature representations can serve as a buffer against catastrophic forgetting. Our findings demonstrate that perfect generalization does not imply equal robustness, offering a new perspective on what makes a trained model robust to interference.

Is Grokking a Computational Glass Relaxation?

Understanding neural network's (NN) generalizability remains a central question in deep learning research. The special phenomenon of grokking, where NNs abruptly generalize long after the training performance reaches a near-perfect level, offers a unique window to investigate the underlying mechanisms of NNs' generalizability. Here we propose an interpretation for grokking by framing it as a computational glass relaxation: viewing NNs as a physical system where parameters are the degrees of freedom and train loss is the system energy, we find memorization process resembles a rapid cooling of liquid into non-equilibrium glassy state at low temperature and the later generalization is like a slow relaxation towards a more stable configuration. This mapping enables us to sample NNs' Boltzmann entropy (states of density) landscape as a function of training loss and test accuracy. Our experiments in transformers on arithmetic tasks suggests that there is NO entropy barrier in the memorization-to-generalization transition of grokking, challenging previous theory that defines grokking as a first-order phase transition. We identify a high-entropy advantage under grokking, an extension of prior work linking entropy to generalizability but much more significant. Inspired by grokking's far-from-equilibrium nature, we develop a toy optimizer WanD based on Wang-landau molecular dynamics, which can eliminate grokking without any constraints and find high-norm generalizing solutions. This provides strictly-defined counterexamples to theory attributing grokking solely to weight norm evolution towards the Goldilocks zone and also suggests new potential ways for optimizer design.

High-entropy Advantage in Neural Networks' Generalizability

While the 2024 Nobel Prize in Physics ignites a worldwide discussion on the origins of neural networks and their foundational links to physics, modern machine learning research predominantly focuses on computational and algorithmic advancements, overlooking a picture of physics. Here we introduce the concept of entropy into neural networks by reconceptualizing them as hypothetical physical systems where each parameter is a non-interacting 'particle' within a one-dimensional space. By employing a Wang-Landau algorithms, we construct the neural networks' (with up to 1 million parameters) entropy landscapes as functions of training loss and test accuracy (or loss) across four distinct machine learning tasks, including arithmetic question, real-world tabular data, image recognition, and language modeling. Our results reveal the existence of \textit{entropy advantage}, where the high-entropy states generally outperform the states reached via classical training optimizer like stochastic gradient descent. We also find this advantage is more pronounced in narrower networks, indicating a need of different training optimizers tailored to different sizes of neural networks.