Researchers set ethics guidance for studies of arXiv sources
They warn public papers can still contain identifiable data, leaked secrets and other sensitive details.
Stories per day · 30 days
3 RSS
Relatedarxiv1data-pipeline1distillation1geolocation1huggingface1imagery1indic1leaked-secrets1multimodal-reasoning1nvidia1post-training1pretraining1project-protostar1reinforcement-learning1research-ethics1satellites1science1tokens1
4 stories, newest first · one card per story
Stories tagged #open-data
Hugging Face announces Project ProtoStar for public science
The announcement says the project will make publicly funded science AI-ready; no further details were available.
MMVistaReason open recipe trains multimodal reasoners with less data
MVR-4B averages 72.8 on 15 benchmarks, beating Qwen3.5-9B Instruct and MMFineReason-8B with about 70% fewer samples.
Scaling Indic Pretraining: 2.8 Trillion Tokens and an Open Data Generation Pipeline
NVIDIA describes scaling Indic pretraining to 2.8 trillion tokens with an open data pipeline.