Which AI stories has SIREN scored in this window? · Which AI news tags have SIREN pages? · Which AI tags has SIREN scored stories about?
What has SIREN scored about Training data?
Training-data and pretraining-corpus stories scored in the newsroom or archive — coverage counts, not a dataset ranking. A tag page lists scored stories whose title or summary matched a published needle — not a probability of harm and not a ranking of the companies named.
Which scored stories are tagged Training data in this window?
- APT: Feedback-Driven Agentic Post-Training for Text-to-Image Generation 47.5
- Binary noise classification using CRNN mean teacher algo [R] 24.1
- Pristine Models 40.9
- Scaling Repeated-Data Pretraining: Batch Size, Regularization, and GPU-Time Tradeoffs 35.3
- Code-Switching Detection via Language-Selective Dimensions in Speech Foundation Models 28.9
- Mix, Don't Tune: Bilingual Pre-Training Outperforms Hyperparameter Search in Data-Constrained Settings 24.7
Which AI stories has SIREN scored in this window?
Which older scored stories are tagged Training data?
Kept in the permanent archive after they left the live window. Still scored stories — not a forecast.
- Elon Musk says acquisition of Cursor is due to the need for programming training data to create superintelligence, with only three institutions having sufficient programming training data 33.1
- DATA Network joins Sui and Walrus to scale the supply chain for AI training data 32.9
- PitchBook: robotics and physical AI companies have raised ~$48B YTD, as they gather training data from people completing tasks in factories, offices, and homes (Rafe Rosner-Uddin/Financial Times) 29.9
- Retrieval-Centric Deep Learning in Growing Nonparametric Neural Networks 34.3
- Toward Joint Optimization of Circuit Depth and Training Data Size in Adaptively Grown Quantum Classifiers 26.8
- NanoProof: Open and Efficient Automated Theorem Proving in Lean 4 27.3
- Adapting English Quality Classifiers for Multilingual LLM Pretraining Data Selection 27.5
- Decoupling Exploration from Optimization in RLVR 27.1
- EmbodiedSmith: Scaling Embodied Data through Recursive Self-Improvement Flywheel in Simulation 39.1
- Source Identification Is Not Fitness Testing: Measuring the Limits of Synthetic-Data Attribution 33.0
- Mistral says it trained ML4 "from scratch" using 3,800 Nvidia Grace Blackwell GPUs in its data centers in Europe and much of its training data was multilingual (Mistral Blog) 34.1
- SAGE: Semantic Anchor-Guided Evolution for Grounded Medical QA Data Synthesis 28.5
Which older AI stories has SIREN kept in the archive?
What questions do people actually ask?
What has SIREN scored about Training data?
40 scored stories match this published needle on SIREN (6 in the live window, 34 older in the archive). Training-data and pretraining-corpus stories scored in the newsroom or archive — coverage counts, not a dataset ranking.
Which scored stories are tagged Training data in this window?
6 scored stories matched this published needle in the current window. A tag page lists headlines — not a topic doom score.
Which older scored stories are tagged Training data?
Kept in the permanent archive after they left the live window. Still scored stories — not a forecast.
Which other tags have scored stories?
Other published tags that clear the substance gate, each linking What has SIREN scored about that tag?
Does SIREN publish a Training data doom score?
No. Activity tempo, not probability of harm. A tag page lists matching headlines — not a probability of harm and not a ranking of the companies named.