Pith. sign in

REVIEW 21 cited by

Transformers Learn Shortcuts to Automata

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2210.10749 v2 pith:I6CXHCOX submitted 2022-10-19 cs.LG cs.FLstat.ML

Transformers Learn Shortcuts to Automata

classification cs.LG cs.FLstat.ML
keywords solutionsmodelsreasoningtransformerautomataautomatoncomputationdepth
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Algorithmic reasoning requires capabilities which are most naturally understood through recurrent models of computation, like the Turing machine. However, Transformer models, while lacking recurrence, are able to perform such reasoning using far fewer layers than the number of reasoning steps. This raises the question: what solutions are learned by these shallow and non-recurrent models? We find that a low-depth Transformer can represent the computations of any finite-state automaton (thus, any bounded-memory algorithm), by hierarchically reparameterizing its recurrent dynamics. Our theoretical results characterize shortcut solutions, whereby a Transformer with $o(T)$ layers can exactly replicate the computation of an automaton on an input sequence of length $T$. We find that polynomial-sized $O(\log T)$-depth solutions always exist; furthermore, $O(1)$-depth simulators are surprisingly common, and can be understood using tools from Krohn-Rhodes theory and circuit complexity. Empirically, we perform synthetic experiments by training Transformers to simulate a wide variety of automata, and show that shortcut solutions can be learned via standard training. We further investigate the brittleness of these solutions and propose potential mitigations.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 21 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Transformers Provably Learn to Internalize Chain-of-Thought

    cs.LG 2026-05 unverdicted novelty 8.0

    L-layer transformers under Log-ICoT curriculum provably learn k-parity with poly(n) samples and log k stages, matching explicit CoT efficiency without inference overhead.

  2. Lost in Tokenization: Fundamental Trade-offs in Graph Tokenization for Transformers

    cs.LG 2026-05 accept novelty 8.0

    Graph tokenizations for Transformers induce distinct depth regimes with proven separations and impossibility results for converting between them at limited depth.

  3. A First-Principles Theory of Slow Thinking and Active Perception

    cs.AI 2026-07 conditional novelty 7.5

    Active lifting of data distributions via latent-sequence sampling and max-rate uncertainty reduction formally derives slow-thinking LLMs and places them on representation and sampler hierarchies that can be climbed.

  4. Structure Before Collapse: Transient semantic geometry in next-token prediction

    cs.LG 2026-06 unverdicted novelty 7.0

    Semantic geometry emerges transiently early in next-token prediction training before collapsing to Neural Collapse symmetry in synthetic settings with latent semantic factors.

  5. Agentic Transformers Provably Learn to Search via Reinforcement Learning

    cs.LG 2026-05 unverdicted novelty 7.0

    In a stochastic k-ary tree, a two-head transformer learns randomized DFS via policy gradient under depth-wise curriculum, generalizes to deeper trees, and adapts to imbalanced goals via discounting.

  6. Do Language Models Need Sleep? Offline Recurrence for Improved Online Inference

    cs.CL 2026-05 unverdicted novelty 7.0

    A sleep mechanism with N offline recurrent passes consolidates context into fast weights, improving performance on reasoning tasks where standard transformers fail.

  7. Learning State-Tracking from Code Using Linear RNNs

    cs.LG 2026-02 unverdicted novelty 7.0

    Linear RNNs track states from REPL code traces of permutations better than Transformers, but non-linear RNNs outperform them in partially observable probabilistic automata.

  8. Scaling Latent Reasoning via Looped Language Models

    cs.CL 2025-10 unverdicted novelty 7.0

    Looped language models with latent iterative computation and entropy-regularized depth allocation achieve performance matching up to 12B standard LLMs through superior knowledge manipulation.

  9. Critical Percolation as a Synthetic Data Model for Interpretability

    cs.LG 2026-06 unverdicted novelty 6.0

    Critical percolation clusters embedded in high dimensions, combined with taxonomic latent variables, form an analytically tractable synthetic data model whose ground-truth hierarchy can be linearly decoded from networ...

  10. Learning Dynamics of Chain-of-Thought State Tracking in a Solvable Transformer Model

    cond-mat.dis-nn 2026-06 unverdicted novelty 6.0

    Mean-field equations for attention retrieval, teacher alignment, and logic overlap quantitatively match simulations and predict a sharp accuracy transition in a solvable transformer for permutation state tracking.

  11. Pretraining Recurrent Networks without Recurrence

    cs.LG 2026-06 unverdicted novelty 6.0

    SMT reduces RNN training to supervised learning on memory transitions (m_t, x_{t+1}) to m_{t+1} obtained from a Transformer encoder, enabling time-parallel training with O(1) gradient paths.

  12. A Close Look At World Model Recovery In Supervised Fine-Tuned LLM Planners

    cs.LG 2026-06 unverdicted novelty 6.0

    Supervised fine-tuning lets LLMs linearly encode action validity and state predicates, with broader state-space coverage during training improving world-model recovery.

  13. A Systematic Study of Behavioral Cloning for Scientific Data Annotation

    cs.HC 2026-05 unverdicted novelty 6.0

    Introduces 9 synthetic annotation tasks and benchmarks for behavioral cloning, finding hierarchical skill learning, scaling benefits, effective multi-task pretraining, and shared internal representations of task phase...

  14. A Sharper Picture of Generalization in Transformers

    cs.LG 2026-05 unverdicted novelty 6.0

    PAC-Bayes applied to low-sharpness flat minima yields non-vacuous generalization bounds for boolean functions whose Fourier spectra are sparse and low-degree, with parameters estimable by property testing.

  15. A Sharper Picture of Generalization in Transformers

    cs.LG 2026-05 unverdicted novelty 6.0

    Sparse low-degree Fourier spectra allow flat minima in transformers for boolean functions up to context-length sparsity, enabling non-vacuous PAC-Bayes generalization bounds via an idealized low-sharpness learner.

  16. Correcting Influence: Unboxing LLM Outputs with Orthogonal Latent Spaces

    cs.LG 2026-05 unverdicted novelty 6.0

    A latent mediation framework with sparse autoencoders enables non-additive token-level influence attribution in LLMs by learning orthogonal features and back-propagating attributions.

  17. The two clocks and the innovation window: When and how generative models learn rules

    cs.LG 2026-05 unverdicted novelty 6.0

    Generative models learn rules before memorizing data, creating an innovation window whose width depends on dataset size and rule complexity, observed in both diffusion and autoregressive architectures.

  18. Do Neural Operators Forget Geometry? The Forgetting Hypothesis in Deep Operator Learning

    cs.LG 2026-05 unverdicted novelty 6.0

    Neural operators progressively forget domain geometry with depth due to Markovian layers and global mixing; a geometry memory injection mechanism mitigates this forgetting.

  19. The Recurrent Transformer: Greater Effective Depth and Efficient Decoding

    cs.LG 2026-04 unverdicted novelty 6.0

    Recurrent Transformers add per-layer recurrent memory via self-attention on own activations plus a tiling algorithm that reduces training memory traffic, yielding better C4 pretraining cross-entropy than parameter-mat...

  20. The Serial Scaling Hypothesis

    cs.LG 2025-07 unverdicted novelty 5.0

    The serial scaling hypothesis formalizes inherently serial problems in complexity theory and demonstrates that diffusion models cannot solve them.

  21. There Will Be a Scientific Theory of Deep Learning

    stat.ML 2026-04 unverdicted novelty 2.0

    A mechanics of the learning process is emerging in deep learning theory, characterized by dynamics, coarse statistics, and falsifiable predictions across idealized settings, limits, laws, hyperparameters, and universa...