• Single MLP Neurons Can Trigger Entity-Level "Amnesia." Google verified causal links across 200 entities — knowledge editing may shift from broad surgery to precision targeting.

• Reusable Problem-Solving Routines Extracted From Reasoning Traces. 32 million procedural knowledge entries free models from reasoning from scratch, with gains up to 19.2%.

• Fine-Tuned LLMs on Real Disruption Events Beat GPT-5 Zero-Shot on accuracy and calibration for supply chain forecasting. The general pattern for low-frequency, high-impact prediction deserves attention.

Also Notable

• Sparse-View 3D Has Kept Reconstruction and Generation Separate — UniRecGen unifies both in a cooperative system, balancing reconstruction fidelity with generative plausibility.
• Bayesian Optimization for Scientific Discovery — a systematic tutorial from fundamentals to practice, for practitioners who want BO to guide experiment design.

Read the full edition →