The paper introduces a compute‑aware three‑action controller for evaluator‑guided reasoning, deciding whether to trust the current answer, fallback, or expand inference. It scores terminal actions by predicted success and uses a rollout‑based surrogate for expansion, explicitly pricing compute cost. Experiments on width expansion and revision tasks show the controller improves utility over fixed policies while reducing online compute by about 30% and 16%, respectively, and demonstrates that evaluator‑guided reasoning is best viewed as resource allocation under uncertainty.
arXiv:2604. 23333v2 Announce Type: replace Abstract: Scaling test-time computation with reinforcement learning (RL) has emerged as a reliable path to improve large language models (LLM) reasoning ability.
arXiv:2605.09860v5 Announce Type: replace Abstract: Long-horizon reasoning requires deciding not only what actions to take, but how many to execute open-loop before replanning. This number, the execu...
arXiv:2606. 06245v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) policies remain brittle in long-horizon and high-uncertainty control, where one-pass action decoding provides limited inference-time deliberation.
The paper introduces CoC-CFG, a method that applies classifier‑free guidance to the reasoning trace of vision‑language‑action (VLA) policies, weighting each Euler step to modulate the influence of reasoning. By treating the weight selection as a Markov decision process with measured action values, the authors develop Q‑adaptive guidance that conservatively improves the controller. Experiments on the Alpasim driving simulator and the Waymo end‑to‑end benchmark show that Q‑adaptive guidance outperforms fixed weights and online RL baselines, achieving a new state‑of‑the‑art score on the hidden test set.
arXiv:2607. 17188v2 Announce Type: replace Abstract: While test-time scaling improves the problem-solving ability of large reasoning models (LRMs) through additional inference-time computation, it can also exacerbate overthinking and underthinking, which we formulate as reasoning state--action mismatch.
arXiv:2607. 02686v1 Announce Type: new Abstract: Reinforcement learning agents operating under partial observability must act on incomplete information, making them natural candidates for guidance from small language models (SLMs) that carry broad reasoning priors.
arXiv:2606. 03965v1 Announce Type: cross Abstract: Large language models improve final-answer accuracy through extended chain-of-thought reasoning, but often spend tokens inefficiently and offer little inference-time control.
arXiv:2610.17309v1 Announce Type: cross Abstract: We introduce StartLux-Decision, decision models from 0.8B to 27B parameters that combine efficient single-pass prediction with adaptive, confidence-g...
The paper proposes a method for selectively querying language‑model advice in reinforcement learning by predicting the value of potential responses and only querying when the expected benefit outweighs the cost. It introduces a certified, response‑contingent metareasoning framework that guarantees near‑optimal advice usage under certain assumptions, and demonstrates that a calibrated controller with Qwen2.5 advisors can improve task performance while drastically reducing the number of advice calls on the BabyAI benchmark.
arXiv:2512.13095v3 Announce Type: replace Abstract: To address the limited capability expansion and low sample efficiency of Reinforcement Learning (RL), recent methods have integrated ''hints'' into...
arXiv:2607. 17240v1 Announce Type: new Abstract: When does a committed intermediate stage in an LLM reasoning pipeline earn its cost?
arXiv:2606. 14581v4 Announce Type: replace Abstract: High-throughput experimentation can evaluate many reaction conditions, yet combinatorial condition spaces still exceed the available experiment budget.