datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
industrial-scada-plc-automation-2026
⚙️ Industrial SCADA & PLC Automation (IEC 61131-3) SFT/DPO Suite
Frontier synthetic alignment dataset engineered for fine-tuning Large Language Models on mission-critical Industrial Automation, Siemens S7 SCL, Rockwell Studio 5000 ST, Beckhoff TwinCAT 3, Schneider M580, IEC 61508 SIL-3 Safety Systems, and SCADA fieldbus telemetry.
📊 Empirical Fine-Tuning Benchmark Delta Matrix
Evaluation Benchmark / Stress Dimension
Base Foundation Model… See the full description on the dataset page: https://huggingface.co/datasets/beatsprom/industrial-scada-plc-automation-2026.ParaDeHate
PARADEHATE
Paper: LLM in the Loop: Creating the ParaDeHate Dataset for Hate Speech Detoxification
Dataset Summary
ParaDeHate is a parallel dataset designed for hate speech detoxification, containing 8,276 pairs of toxic (hate speech) and detoxified (non-hateful) text samples. It was created using an LLM-in-the-loop pipeline with GPT-4o-mini, automating the process of rewriting hate speech into non-toxic, semantically equivalent text. The dataset is sourced from four… See the full description on the dataset page: https://huggingface.co/datasets/ScaDSAI/ParaDeHate.