Research
DeepMind publishes new results on efficient sparse training for large models
New techniques aim to reduce training cost while preserving downstream quality across reasoning and coding evaluations.
Research Desk — 15 days ago
ML
Research, arXiv, evaluation, and model releases.
ModelsML Engineering — 16 days ago
Meta releases Llama‑4 with enhanced reasoning capabilities
The release tightens the open model race, with improved tool-use and stronger performance in long-context tasks.
arXivResearch Desk — 17 days ago
arXiv Weekly: RLHF evaluation is shifting from preference to reliability
New work suggests preference tuning must be paired with robustness metrics if models are to operate safely as agents.