Entropy-Gated Latent Recursion

• URL: http://arxiv.org/abs/2606.16620v1
• Date: Mon, 15 Jun 2026 12:14:01 GMT
• Title: Entropy-Gated Latent Recursion
• Abstract summary: In-time scaling has become the dominant lever for improving language-model reasoning.<n>Existing methods derive rollout diversity from a single source: token-level sampling.<n>We argue that this single-axis sampling space is fundamentally limiting.<n>We show that the $L$-axis is genuinely complementary to temperature.
• Score: 9.65821666936513
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Inference-time scaling has become the dominant lever for improving language-model reasoning, but existing methods derive rollout diversity from a single source: stochastic token-level sampling. We argue that this single-axis sampling space is fundamentally limiting, and identify a second, fully deterministic and complementary axis: the layer span $L$ at which a frozen model's top decoder layers are recursively re-applied at high-uncertainty tokens. Different choices of $L$ produce distinct rollouts that solve different subsets of problems, with no stochasticity. We instantiate this axis through Entropy-Gated Latent Recursion (EGLR), a training-free decoding procedure that re-applies the top-$L$ layers for at most $K_{\max}$ iterations until the next-token distribution converges. Combined with $T$ temperature samples, EGLR turns a single-axis stochastic rollout pool into an $L\times T$ Cartesian sampling space at almost the same per-rollout cost. We characterize this space across $8$ instruction-tuned models and $6$ math reasoning benchmarks, and show that the $L$-axis is genuinely complementary to temperature: on MATH-500 with Qwen2.5-3B-Instruct, the joint $L\times T$ oracle reaches $91.6\%$, $+8.2$ percentage points beyond the temperature-only oracle ($83.4\%$) and $+10.4$ points beyond the layer-only oracle ($81.2\%$), confirming that the two axes capture genuinely complementary problems. The expanded rollout pool provides richer per-prompt candidates for any downstream procedure that consumes rollouts, including self-consistency, best-of-$N$ with verifiers, and group-relative RL training (GRPO), opening a new direction for inference-time scaling that does not rely on stochastic noise.

Related papers

• Round-Trip Consistency: Bidirectional Diffusion Models Can Predict Their Own Rollout Errors [51.56484100374058]
  Autoregressive models accumulate error over long rollouts, yet at deployment there is no ground truth to measure it against.<n>We train a single conditional latent diffusion model that steps a dynamical system forward or backward in time via a direction flag.<n>We show that this bidirectionality supplies a measurement-free test-time error signal.
  arXiv  Detail & Related papers  (2026-08-01T13:49:46Z)
• Minimax PAC Bounds for Learning in Exogenous Contextual MDPs [26.599514971653615]
  PAC learning in discounted decision processes with i.i.d. contexts, with discount factor $$, finite state space $mathcal X$, action space $mathcal A$, and context space $mathcal Z$.
  arXiv  Detail & Related papers  (2026-06-23T21:02:18Z)
• SILAGE: Memory-Efficient, Full-Gradient-Free Nonconvex Optimization for Nested Finite Sums [51.49970814177172]
  Empirical risk on datasets naturally exhibits similarities in $N=nm$ total samples.<n>We provide an analysis that avoids pessimistic convergence analysis.<n>Our results improve existing state-of-the-art regimes.
  arXiv  Detail & Related papers  (2026-06-14T14:11:07Z)
• Sequential Physics-Constrained Neural Operator Forward Modeling for the $\ extit{Norne}$ Reservoir System [3.175986877079391]
  We develop a comprehensive framework for sequential surrogate modeling of black-oil reservoir dynamics using neural operators.<n>The application focus is the Norne benchmark reservoir, defined on a heterogeneous $46times112times22$ grid.<n>A 1000-member ensemble runs in under one minute on a single B200 GPU, giving a $sim104times$ wall-clock speedup over the OPM finite-volume simulator.
  arXiv  Detail & Related papers  (2026-05-27T17:18:39Z)
• Scalable Maximum Entropy Population Synthesis via Persistent Contrastive Divergence [0.0]
  A principled framework for generating synthetic populations from aggregate census data is presented.<n>GibbsPCDr produces populations with effective sample size versus $Neff approx 0.012, raking an $86.8times$ diversity advantage that is essential for agent-based urban simulations.<n>GibbsPCDr maintains $MRE in [0.010, 0.018]$ while $|cX|$ grows eighteen orders of magnitude, with scaling as $O(K)$ rather than $O(cX)$.
  arXiv  Detail & Related papers  (2026-03-28T15:36:20Z)
• Parallel Sampling via Autospeculation [13.643401888306398]
  We present algorithms to accelerate sampling via counting in two settings: any-order autoregressive models and denoising diffusion models.<n>We show that, by issuing oracle calls in parallel, the expected sampling time can be reduced to $widetildeO(n1/2)$.<n>We introduce a novel technique to obtain our results: speculative rejection sampling.
  arXiv  Detail & Related papers  (2025-11-11T06:09:44Z)
• From Continual Learning to SGD and Back: Better Rates for Continual Linear Models [50.11453013647086]
  We analyze the forgetting, i.e., loss on previously seen tasks, after $k$ iterations.<n>We develop novel last-iterate upper bounds in the realizable least squares setup.<n>We prove for the first time that randomization alone, with no task repetition, can prevent catastrophic in sufficiently long task sequences.
  arXiv  Detail & Related papers  (2025-04-06T18:39:45Z)
• Stochastic Approximation Approaches to Group Distributionally Robust Optimization and Beyond [89.72693227960274]
  This paper investigates group distributionally robust optimization (GDRO) with the goal of learning a model that performs well over $m$ different distributions. To reduce the number of samples in each round from $m$ to 1, we cast GDRO as a two-player game, where one player conducts and the other executes an online algorithm for non-oblivious multi-armed bandits. In the second scenario, we propose to optimize the average top-$k$ risk instead of the maximum risk, thereby mitigating the impact of distributions.
  arXiv  Detail & Related papers  (2023-02-18T09:24:15Z)
• Reward-Mixing MDPs with a Few Latent Contexts are Learnable [75.17357040707347]
  We consider episodic reinforcement learning in reward-mixing Markov decision processes (RMMDPs) Our goal is to learn a near-optimal policy that nearly maximizes the $H$ time-step cumulative rewards in such a model.
  arXiv  Detail & Related papers  (2022-10-05T22:52:00Z)
• A Variance-Reduced Stochastic Accelerated Primal Dual Algorithm [3.2958527541557525]
  Such problems arise frequently in machine learning in the context of robust empirical risk minimization. We consider the accelerated primal dual (SAPD) algorithm as a robust method against gradient noise. We show that our method improves upon SAPD both in practice and in theory.
  arXiv  Detail & Related papers  (2022-02-19T22:12:30Z)
• Reward-Free Model-Based Reinforcement Learning with Linear Function Approximation [92.99933928528797]
  We study the model-based reward-free reinforcement learning with linear function approximation for episodic Markov decision processes (MDPs) In the planning phase, the agent is given a specific reward function and uses samples collected from the exploration phase to learn a good policy. We show that to obtain an $epsilon$-optimal policy for arbitrary reward function, UCRL-RFE needs to sample at most $tilde O(H4d(H + d)epsilon-2)$ episodes.
  arXiv  Detail & Related papers  (2021-10-12T23:03:58Z)
• Sample Complexity of Asynchronous Q-Learning: Sharper Analysis and Variance Reduction [63.41789556777387]
  Asynchronous Q-learning aims to learn the optimal action-value function (or Q-function) of a Markov decision process (MDP) We show that the number of samples needed to yield an entrywise $varepsilon$-accurate estimate of the Q-function is at most on the order of $frac1mu_min (1-gamma)5varepsilon2+ fract_mixmu_min (1-gamma)$ up to some logarithmic factor.
  arXiv  Detail & Related papers  (2020-06-04T17:51:00Z)

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.