CORE: Contrastive Reflection Enables Rapid Improvements in Reasoning
Paper • 2605.28742 • Published • 4
Insights circa Solstice 2026
-
-
Reinforcement Learning from Rich Feedback with Distributional DAgger
Paper • 2606.05152 • Published • 3 -
Entropy as a Structural Prior: How a Log-Barrier on DiT Belief Space Drives Musical Diversity and Development
Paper • 2606.07207 • Published • 4 -
Bayesian-Agent: Posterior-Guided Skill Evolution for LLM Agent Harnesses
Paper • 2606.08348 • Published • 14 -
FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse Attention
Paper • 2606.09079 • Published • 61 -
UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs
Paper • 2606.06622 • Published • 20 -
LLM Explainability with Counterfactual Chains and Causal Graphs
Paper • 2606.05972 • Published • 16 -
Echo-Memory: A Controlled Study of Memory in Action World Models
Paper • 2606.09803 • Published • 32 -
When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents
Paper • 2606.05806 • Published • 22 -
Human Psychometric Questionnaires Mischaracterize LLM Behavior
Paper • 2509.10078 • Published • 35 -
OpenSkill: Open-World Self-Evolution for LLM Agents
Paper • 2606.06741 • Published • 27 -
Direct 3D-Aware Object Insertion via Decomposed Visual Proxies
Paper • 2606.06601 • Published • 26 -
Robots Need More than VLA and World Models
Paper • 2606.06556 • Published • 27 -
AnchorWorld: Embodied Egocentric World Simulation with View-based Evolution Customization
Paper • 2606.07326 • Published • 29 -
On the Geometry of On-Policy Distillation
Paper • 2606.07082 • Published • 71 -
SoCRATES: Towards Reliable Automated Evaluation of Proactive LLM Mediation across Domains and Socio-cognitive Variations
Paper • 2606.05563 • Published • 50 -
Your UnEmbedding Matrix is Secretly a Feature Lens for Text Embeddings
Paper • 2606.07502 • Published • 91 -
When Gradients Collide: Failure Modes of Multi-Objective Prompt Optimization for LLM Judges
Paper • 2605.26046 • Published • 3