The paper introduces the Communication Map, a method that charts every potential communication channel in a transformer model using only its weights. It generalizes previous coupling metrics into a single coefficient covering all 18 connection classes, revealing that 70‑89% of head pairs are non‑randomly oriented and identifying strong or avoiding couplings. The authors demonstrate the map’s utility by recovering known induction circuits and uncovering a two‑dimensional stream subspace whose removal eliminates induction capabilities across several models.
By Richard Zhe Wang
arXiv:2606. 19150v1 Announce Type: new Abstract: The remarkable success of Transformer-based models in natural language processing stems from architectural scaling, which leads to a large number of parameters and hinders deployment in resource-constrained environments.
By Yaniv Livertovsky, Shahar Somin, Gonen Singer
arXiv:2601.06787v2 Announce Type: replace
Abstract: Large Language Models (LLMs) are known to contain significant redundancy, yet a systematic explanation for why certain components, particularly in...
By Jaewon Sok, Jewon Yeom, Seonghyeon Park, Jeongjae Park, Taesup Kim
arXiv:2607. 01940v1 Announce Type: cross Abstract: Mechanistic interpretability often relies on component-level interventions to discover how a model produces a behavior.
By Zhiren Gong, Zihao Zeng, Chau Yuen, Wei Yang Bryan Lim
The paper investigates how knowledge distillation affects the internal mechanisms of language models by comparing GPT‑2 with its six‑layer student, DistilGPT‑2. Using techniques such as linear CKA, logit lens analysis, and attention pattern decomposition, the authors examine two circuits: Indirect Object Identification (IOI) and Greater‑Than (GT). They find that the IOI circuit collapses completely in the distilled model, whereas the GT circuit largely survives, with the student’s layer‑4 head effectively remapping the teacher’s layer‑9 comparison head, achieving 87.7% of the original task performance.
By Nabeelah Faumi, Deshan Sumanathilaka
Sparse autoencoders (SAEs) are commonly used to interpret large language models, but their reliability after pruning is unclear. This study shows that pruning’s effect on an SAE is governed by perturbation energy, a covariance-weighted norm, and that magnitude pruning distorts the representation space by ignoring activation geometry. Activation-aware pruning methods such as Wanda and SparseGPT better preserve SAE behavior, and the authors find that middle layers are especially vulnerable, leading them to propose a layer‑wise sparsity allocation that reduces perplexity for a given sparsity level.
By Suchit Gupte, Xueru Zhang, Mohammad Mahdi Khalili
arXiv:2602. 22600v2 Announce Type: replace-cross Abstract: Training selects for behavior, not circuitry: many weight configurations can implement the same function.
By Joshua S. Schiffman
arXiv:2606. 30815v1 Announce Type: cross Abstract: Recent work suggests that transformer language models show a bias towards human languages over unnatural ("impossible") languages argued to be unacquirable by humans.
By Ram Janarthan, Coleman Haley, Sharon Goldwater
Magnitude Profile Pruning introduces a training‑free, calibration‑free method for removing attention heads in Transformer models by statistically detecting outliers in weight row norms. Heads whose projection weights fall within the bulk of the distribution are pruned, while outlier heads are retained. Across several models, the MP‑G variant achieves superior perplexity at various sparsity levels and yields significant parameter and FLOP reductions without requiring forward passes, calibration data, or gradient computations.
By Kasun Dewage, Marianna Pensky, Heranga K. Rathnasekara, Suranadi De Silva
The paper studies GLU-based neurons in large language models by measuring the cosine similarity between each neuron's input and output weight vectors. A strong negative similarity identifies a "weakening neuron," which tends to appear in late layers, activates frequently, and exerts a large influence on model behavior. The authors also find that weakening neurons significantly affect outputs when gate values are negative, contrary to expectations.
By Sebastian Gerstner, Hilal AlQuabeh, Kentaro Inui, Hinrich Sch\"utze
The paper presents a retraining‑free technique that transforms pre‑trained transformers into elastic models capable of adapting to any computational budget on the fly. By jointly ranking every MLP neuron and attention head on a common scale and applying soft pruning across multiple sparsity targets, the method achieves smooth performance degradation up to 60% sparsity while outperforming one‑shot baselines. Experiments on 27 vision and text transformer encoders show significant parameter reductions with minimal accuracy loss and notable throughput and latency gains on GPUs, all produced in minutes.
By Walter Simoncini, Dheeraj Varghese, Gertjan J. Burghouts, Cees G. M. Snoek
arXiv:2606. 30676v1 Announce Type: cross Abstract: Deploying spiking neural networks (SNNs) on neuromorphic hardware demands aggressive synaptic pruning while preserving temporal computation integrity.
By Muhammad Hamza