The paper introduces entry-state sharpening, a data‑free pre‑training step that prepares a language model’s checkpoint in a sharper, lower‑entropy state before test‑time reinforcement learning (TTRL). By reducing policy entropy, the model can more efficiently use its limited adaptation budget, leading to higher endpoint conversion efficiency across tasks such as MATH, GPQA, and AMC. Experiments with different data‑free objectives (e.g., R‑Zero vs. SPIRAL) demonstrate that the choice of pre‑training objective strongly influences the checkpoint’s readiness for TTRL, and a label‑free self‑distillation intervention can further sharpen the entry state.
arXiv:2609.06107v1 Announce Type: new Abstract: Data policies for reinforcement learning with verifiable rewards (RLVR) determine which rollouts are used, how strongly they are weighted, and which do...
The paper introduces temporal self‑distillation for reinforcement learning with verifiable rewards (RLVR), proposing that a policy can learn from a stronger future checkpoint of itself. Two methods—Near‑Future Policy Optimization (NPO) and Near‑Future Policy Distillation (NPD)—use verified future‑self trajectories and token‑level transfer, respectively, while AutoNPO adaptively selects the optimal future checkpoint. Experiments on eight image‑text benchmarks show that near‑future teachers yield higher performance than far‑future ones, indicating that the balance between new capability and learner compatibility is key.
The paper introduces the concept of intervention fidelity in latent world models, measuring whether a model’s open‑loop transitions align with actual environment interventions. Experiments on TD‑MPC2, Cheetah, and DreamerV3 show that high reward fit does not guarantee fidelity, and that self‑supervised models can outperform task‑anchored ones in preserving intervention effects. The authors propose a capture‑gated audit to localize failures and argue that fidelity must be directly audited on the model’s native interface.
arXiv:2609.35954v1 Announce Type: cross Abstract: Large language model post-training generates self-generated rollouts through reinforcement learning and on-policy distillation, yet this experience i...
The paper investigates why reinforcement learning with verifiable rewards (RLVR) reduces the diversity of solutions in reasoning tasks. By analyzing the Countdown task, the authors show that RLVR contracts the solution space mainly at the entrance—before the first arithmetic operation—causing a 67% drop in solution coverage. They demonstrate that providing an unselected entrance prefix or applying entrance‑targeted interventions can restore or even improve coverage without harming accuracy.
arXiv:2606.28876v4 Announce Type: replace-cross Abstract: Memory-Mediated Learning Architecture (MMLA) separates slow base parameters theta, a bounded numerical policy carrier Phi, and a bounded auth...
arXiv:2609.36569v1 Announce Type: cross Abstract: Checkpoint selection is a routine decision in supervised fine-tuning (SFT): training produces multiple checkpoints, but only one is retained. Yet fix...
arXiv:2607. 28457v1 Announce Type: cross Abstract: Scaling test-time computation can improve language-model reasoning, but uniform budgets waste computation on easy inputs, while verifier-guided refinement relies on external feedback.
arXiv:2609. 03241v1 Announce Type: cross Abstract: A reasoning model can improve from its own on-policy experience, but this inner loop is fragile: terminal verifiers provide reliable yet sparse supervision, while dense same-model guidance can reinforce false confidence or overconcentrate learning on a narrow solution mode.
ComputerSD is an online self‑distillation method for computer‑use agents that leverages real‑time feedback from executed GUI transitions. It uses a fine‑tuned GUI analyzer to generate guidance and a step‑level value score after each action, combining token‑level OPSD with trajectory‑level GRPO in an asynchronous training framework. On the OSWorld‑Verified benchmark, ComputerSD improves performance over outcome‑only GRPO by 1.9 and 4.1 percentage points on Qwen3‑VL‑8B‑Thinking and EvoCUA‑8B backbones, and shows strong generalizability in out‑of‑distribution tests.
arXiv:2609.38142v1 Announce Type: new Abstract: A small trainable advisor can steer a frozen language-model executor using natural-language advice. In addition to learning from task rewards, the advi...
PROOF-Gen is a method that improves distillation of tool‑calling models by recovering successful trajectories from teacher failures. It uses per‑scenario prompt optimization to generate corrective guidance that steers the teacher to a passing trajectory, then removes this guidance before training so the student learns from clean demonstrations. On τ2‑bench, PROOF-Gen recovers 93% of failed scenarios, boosting Qwen3‑4B‑Instruct‑2507’s Pass^1 from 0.132 to 0.529 and improving Gemma 4 E4B‑it by 7.2pp on BFCL v4 multi‑turn, while also raising deployed on‑device model performance by up to 5.0pp across response‑quality metrics.