geodesic-research/inoculation-midtraining-mixes download history

geodesic-research/inoculation-midtraining-mixes is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 1,207 times (17 in the last 7 days), and 9,853 times in total. It ranks #16,773 among datasets by monthly downloads.

Inoculation Midtraining Mixes Synthetic training data for AI safety research exploring how language models respond to stage-awareness tags (<stage=training>, <stage=deployment>). All data was generated using vLLM batch inference on the Isambard AI supercomputer with NousResearch/Hermes-4-

Open geodesic-research/inoculation-midtraining-mixes on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.