Agon: Competitive Cross-Model RL with Implicit Rival Grading of Reasoning
arXiv:2607. 07690v1 Announce Type: cross Abstract: Reinforcement learning from verifiable rewards (e.
arXiv:2607. 07690v1 Announce Type: cross Abstract: Reinforcement learning from verifiable rewards (e.
The paper demonstrates that language models can acquire new capabilities from post‑training data even when the training text is unrelated to the target task. Using a method called Active Taskless Distillation (ATD), the authors show that a single word from a teacher model can transfer knowledge to a student model without any target‑task examples or teacher logits. Experiments on Qwen2.5-1.5B reveal significant performance gains on HumanEval+ and improvements in scientific knowledge, commonsense reasoning, and reading comprehension across various model families.
The paper investigates whether groups of neurons that influence a model’s answer are uniquely necessary and transferable across inputs. By exhaustively testing all 4,096 subsets of 12 MLP neurons in Qwen3 and other models, the authors find that only a few neurons are often sufficient to alter predictions, that many subsets can independently flip answers, and that the effect is strongest on the same input chain but weaker on different chains. The study shows that causal neuron circuits are not uniquely defined and that their influence can vary with context and model architecture.
The paper investigates the problem of sharing a single critic across multiple parallel environments in reinforcement learning. It shows that when environments assign different expected returns to the same state, a shared critic must reconcile conflicting value targets, which can distort advantage estimates and misguide policy updates. The authors propose a simple fix—providing the critic with the environment index—demonstrating through bandit models and experiments on CartPole, MuJoCo, BipedalWalker, and 16 Procgen games that this conditional critic stabilizes learning and boosts returns, achieving a 40.8% improvement in aggregate normalized return on unseen levels.
The paper reviews On‑Policy Self‑Distillation (OPSD), a method where a language model learns from its own generations using privileged information such as reference solutions or plans, eliminating the need for a larger teacher model. It identifies a key failure mode—collapse, where the model’s reasoning paths narrow progressively—and analyzes it through three levers: signal application, privileged information, and teacher dynamics. The review focuses on mathematical reasoning, offering a unified vocabulary and distinguishing settled facts from ongoing debates.
The paper investigates whether large language models (LLMs) can learn from rewards presented in context, a process termed in-context reinforcement learning (ICRL). Experiments across four benchmarks and six models show that while LLMs read reward signals, altering or removing these rewards has little impact on performance, and shuffled or corrupted trajectories are as effective as real ones. These findings suggest that direct ICRL behaves more like in-context learning (ICL) than true inference-time reinforcement learning.
arXiv:2607. 25063v1 Announce Type: new Abstract: Developers judge a model checkpoint by how it behaves.
arXiv:2607. 01480v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR), along with recent selfdistillation variants such as SDPO, evaluates each rollout against a verifier and updates the policy from that episode-level signal.
arXiv:2608. 05111v1 Announce Type: new Abstract: In partially observable reinforcement learning, agents face a dual bottleneck: they must explore to encounter rewarding states and retain that experience in memory to optimize their policies.
arXiv:2601. 17454v2 Announce Type: replace-cross Abstract: Centralized value learning underlies a broad class of multi-agent reinforcement learning methods, but its claimed advantage is typically evaluated in settings that confound coordination structure with function approximation and partial observability.
CausalDreamer enhances generative world models by freezing the original video tokenizer and re-encoding its latent space into a factored representation that separates controllable versus uncontrollable dynamics and reward-relevant versus reward-irrelevant information. The pretrained dynamics model is fine‑tuned to predict this factored representation, enabling more accurate model‑predictive planning. On the MMBench2 benchmark, CausalDreamer improves normalized scores by 14% on clean tasks and 25% on manipulated variants compared to the baseline, while both models perform poorly on entirely new environments.
Probes on activation vectors are widely used to detect misbehavior in language models, but their success relies on the assumption that models cannot control what their activations encode. Humans can l...