We develop a unified mathematical framework for certified Top-$k$ attention truncation that quantifies approximation error at both the distribution and output levels. For a single attention distribution $P$ and its Top-$k$ truncation $\hat P$, we show that the total-variation distance coincides with the discarded softmax tail mass and satisfies $\mathrm{TV}(P,\hat P)=1-e^{-\mathrm{KL}(\hat P\Vert P)}$, yielding sharp Top-$k$-specific bounds in place of generic inequalities. From this we derive non-asymptotic deterministic bounds -- from a single boundary gap through multi-gap and blockwise variants -- that control $\mathrm{TV}(P,\hat P)$ using only the ordered logits. Using an exact head-tail decomposition, we prove that the output error factorizes as $\|\mathrm{Attn}(q,K,V)-\mathrm{Attn}_k(q,K,V)\|_2=τ\|μ_{\mathrm{tail}}-μ_{\mathrm{head}}\|_2$ with $τ=\mathrm{TV}(P,\hat P)$, yielding a new head-tail diameter bound $\|\mathrm{Attn}(q,K,V)-\mathrm{Attn}_k(q,K,V)\|_2\leτ\,\mathrm{diam}_{H,T}$ and refinements linking the error to $\mathrm{Var}_P(V)$. Under an i.i.d. Gaussian score model $s_i\sim\mathcal N(μ,σ^2)$ we derive closed-form tail masses and an asymptotic rule for the minimal $k_\varepsilon$ ensuring $\mathrm{TV}(P,\hat P)\le\varepsilon$, namely $k_\varepsilon/n\approxΦ_c(σ+Φ^{-1}(\varepsilon))$. Experiments on bert-base-uncased and synthetic logits confirm the predicted scaling of $k_\varepsilon/n$ and show that certified Top-$k$ can reduce scored keys by 2-4$\times$ on average while meeting the prescribed total-variation budget.

Finding human-understandable circuits in language models is a central goal of the field of mechanistic interpretability. We train models to have more understandable circuits by constraining most of their weights to be zeros, so that each neuron only has a few connections. To recover fine-grained circuits underlying each of several hand-crafted tasks, we prune the models to isolate the part responsible for the task. These circuits often contain neurons and residual channels that correspond to natural concepts, with a small number of straightforwardly interpretable connections between them. We study how these models scale and find that making weights sparser trades off capability for interpretability, and scaling model size improves the capability-interpretability frontier. However, scaling sparse models beyond tens of millions of nonzero parameters while preserving interpretability remains a challenge. In addition to training weight-sparse models de novo, we show preliminary results suggesting our method can also be adapted to explain existing dense models. Our work produces circuits that achieve an unprecedented level of human understandability and validates them with considerable rigor.

The number of layers is 12 for GPT2 and randomly initialized model and 24 for iGPT. Note that these notations are sometimes used interchangeably as long as it doesn't significantly The activation to be analyzed are outputs from all layers . CKA about is shown in Figure 1. The design of the diagram is based on a previous study [35]. Figure 11: Activation we consider to compute CKA.

Our study reveals the roles of critical parameters in the Transformer, such as the number of layers and the number of attention heads.

Guo. Swin transformer: Hierarchical vision transformer using shifted windows.

The ability to reason lies at the core of artificial intelligence (AI), and challenging problems usually call for deeper and longer reasoning to tackle. A crucial question about AI reasoning is whether models can extrapolate learned reasoning patterns to solve harder tasks with longer chain-of-thought (CoT). In this work, we present a theoretical analysis of transformers learning on synthetic state-tracking tasks with gradient descent. We mathematically prove how the algebraic structure of state-tracking problems governs the degree of extrapolation of the learned CoT. Specifically, our theory characterizes the length generalization of transformers through the mechanism of attention concentration, linking the retrieval robustness of the attention layer to the state-tracking task structure of long-context reasoning. Moreover, for transformers with limited reasoning length, we prove that a recursive self-training scheme can progressively extend the range of solvable problem lengths. To our knowledge, we provide the first optimization guarantee that constant-depth transformers provably learn $\mathsf{NC}^1$-complete problems with CoT, significantly going beyond prior art confined in $\mathsf{TC}^0$, unless the widely held conjecture $\mathsf{TC}^0 \neq \mathsf{NC}^1$ fails. Finally, we present a broad set of experiments supporting our theoretical results, confirming the length generalization behaviors and the mechanism of attention concentration.

We introduce TwIST, a distributed training framework for efficient large language model (LLM) sparsification. TwIST trains multiple subnetworks in parallel, periodically aggregates their parameters, and resamples new subnetworks during training. This process identifies high-quality subnetworks ("golden tickets") without requiring post-training procedures such as calibration or Hessian-based recovery. As a result, TwIST enables zero-cost pruning at deployment time while achieving perplexity competitive with state-of-the-art post-training sparsification methods. The benefits are most pronounced under aggressive sparsity (e.g., 50%+), where TwIST significantly outperforms baseline methods; for example, reaching 23.14 PPL compared to 31.64 for the closest prior approach. Unlike unstructured pruning, TwIST produces structured, dense matrices that offer practical inference speedups and memory reductions on commodity hardware (e.g., CPUs) that do not support efficient sparse computation. TwIST provides an efficient training-time path to deployable sparse LLMs without additional fine-tuning or recovery overhead.