Loop, Think, & Generalize: Implicit Reasoning in Recurrent-Depth Transformers

• URL: http://arxiv.org/abs/2604.07822v1
• Date: Thu, 09 Apr 2026 05:24:32 GMT
• Title: Loop, Think, & Generalize: Implicit Reasoning in Recurrent-Depth Transformers
• Abstract summary: We study implicit reasoning, i.e. the ability to combine knowledge or rules within a single forward pass.<n>While transformer-based large language models store substantial factual knowledge and rules, they often fail to compose this knowledge for implicit multi-hop reasoning.<n>We show that while vanilla transformers struggle with both generalization challenges, recurrent-depth transformers can effectively make such generalization.
• Score: 14.641203402842672
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We study implicit reasoning, i.e. the ability to combine knowledge or rules within a single forward pass. While transformer-based large language models store substantial factual knowledge and rules, they often fail to compose this knowledge for implicit multi-hop reasoning, suggesting a lack of compositional generalization over their parametric knowledge. To address this limitation, we study recurrent-depth transformers, which enables iterative computation over the same transformer layers. We investigate two compositional generalization challenges under the implicit reasoning scenario: systematic generalization, i.e. combining knowledge that is never used for compositions during training, and depth extrapolation, i.e. generalizing from limited reasoning depth (e.g. training on up to 5-hop) to deeper compositions (e.g. 10-hop). Through controlled studies with models trained from scratch, we show that while vanilla transformers struggle with both generalization challenges, recurrent-depth transformers can effectively make such generalization. For systematic generalization, we find that this ability emerges through a three-stage grokking process, transitioning from memorization to in-distribution generalization and finally to systematic generalization, supported by mechanistic analysis. For depth extrapolation, we show that generalization beyond training depth can be unlocked by scaling inference-time recurrence, with more iterations enabling deeper reasoning. We further study how training strategies affect extrapolation, providing guidance on training recurrent-depth transformers, and identify a key limitation, overthinking, where excessive recurrence degrades predictions and limits generalization to very deep compositions.

Related papers

• Thinking Deeper, Not Longer: Depth-Recurrent Transformers for Compositional Generalization [1.5736899098702974]
  We propose a depth-recurrent Transformer that decouples computational depth from parameter count.<n>Our architecture incorporates three mechanisms to make deep recurrence (20+ steps) stable.<n>We observe a clear emphcomputational frontier, where performance transitions from chance to near-perfect as thinking steps scale with task complexity.
  arXiv  Detail & Related papers  (2026-03-23T08:06:45Z)
• Is Grokking Worthwhile? Functional Analysis and Transferability of Generalization Circuits in Transformers [15.965423731432422]
  We conduct a study to evaluate the Generalization Circuit's role in knowledge assimilation and transfer.<n>We argue that grokking is the process of integrating memorized atomic facts into an naturally established reasoning path.
  arXiv  Detail & Related papers  (2026-01-14T00:40:35Z)
• Transformers Provably Learn Chain-of-Thought Reasoning with Length Generalization [53.89723291716722]
  A crucial question about AI reasoning is whether models can extrapolate learned reasoning patterns to solve harder tasks with longer chain-of-thought (CoT)<n>We mathematically prove how the algebraic structure of state-tracking problems governs the degree of extrapolation of the learned CoT.<n>We provide the first optimization guarantee that constant-depth transformers provably learn $mathsfNC1$-complete problems with CoT.
  arXiv  Detail & Related papers  (2025-11-10T18:40:24Z)
• Quantitative Bounds for Length Generalization in Transformers [58.175107357008876]
  We study the problem of length generalization (LG) in transformers.<n>LG occurs when the internal behavior of the transformer on longer sequences can be "simulated" by its behavior on shorter sequences.
  arXiv  Detail & Related papers  (2025-10-30T21:31:36Z)
• Multi-head Transformers Provably Learn Symbolic Multi-step Reasoning via Gradient Descent [66.78052387054593]
  This work investigates how transformers learn to solve symbolic multi-step reasoning problems through chain-of-thought processes.<n>We analyze two intertwined tasks: a backward reasoning task, where the model outputs a path from a goal node to the root, and a more complex forward reasoning task.<n>We show that trained one-layer transformers can provably solve both tasks with generalization guarantees to unseen trees.
  arXiv  Detail & Related papers  (2025-08-11T17:40:47Z)
• Extrapolation by Association: Length Generalization Transfer in Transformers [29.659527141850436]
  We show that length generalization can be textittransferred across related tasks.<n>Our results show that transformer models can inherit generalization capabilities from similar tasks when trained jointly.<n>We provide initial mechanistic evidence that length generalization transfer correlates with the re-use of the same attention heads between the tasks.
  arXiv  Detail & Related papers  (2025-06-10T21:22:51Z)
• How do Transformers Learn Implicit Reasoning? [67.02072851088637]
  We study how implicit multi-hop reasoning emerges by training transformers from scratch in a controlled symbolic environment.<n>We find that training with atomic triples is not necessary but accelerates learning, and that second-hop generalization relies on query-level exposure to specific compositional structures.
  arXiv  Detail & Related papers  (2025-05-29T17:02:49Z)
• Interpreting Affine Recurrence Learning in GPT-style Transformers [54.01174470722201]
  In-context learning allows GPT-style transformers to generalize during inference without modifying their weights. This paper focuses specifically on their ability to learn and predict affine recurrences as an ICL task. We analyze the model's internal operations using both empirical and theoretical approaches.
  arXiv  Detail & Related papers  (2024-10-22T21:30:01Z)
• A Formal Framework for Understanding Length Generalization in Transformers [14.15513446489798]
  We introduce a rigorous theoretical framework to analyze length generalization in causal transformers.<n>We experimentally validate the theory as a predictor of success and failure of length generalization across a range of algorithmic and formal language tasks.
  arXiv  Detail & Related papers  (2024-10-03T01:52:01Z)
• In-Context Learning with Representations: Contextual Generalization of Trained Transformers [66.78052387054593]
  In-context learning (ICL) refers to a capability of pretrained large language models, which can learn a new task given a few examples during inference. This paper investigates the training dynamics of transformers by gradient descent through the lens of non-linear regression tasks.
  arXiv  Detail & Related papers  (2024-08-19T16:47:46Z)
• Grokked Transformers are Implicit Reasoners: A Mechanistic Journey to the Edge of Generalization [22.033370572209744]
  We study whether transformers can learn to implicitly reason over parametric knowledge. We focus on two representative reasoning types, composition and comparison. We find that transformers can learn implicit reasoning, but only through grokking.
  arXiv  Detail & Related papers  (2024-05-23T21:42:19Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.