Alignment datasets by EleutherAI

Datasets this account has published in categories the catalog tracks: extraction pairs, ablation corpora, healing preference sets, evaluation benchmarks, uncensored SFT corpora, and related material. Category badges link to the workflow stage.

4 in /datasets
Dataset
Stage
Downloads
fake-mnist 0 0
This is a dataset of "fake" MNIST images which were sampled from a high-entropy distribution whose mean and covariance matrix matches that o...
747
fake-cifar10 0 0
This is a dataset of "fake" CIFAR-10 images which were sampled from a high-entropy distribution whose mean and covariance matrix matches tha...
651
fake-cifarnet 0 0
This is a dataset of "fake" CIFARNet images which were sampled from a high-entropy distribution whose mean and covariance matrix matches tha...
582
rh-clean-control-sft 0 0
Clean Control SFT Mixture A clean SFT mixture dataset for use as a control in reward hacking experiments. This dataset contains only beni...
451
Dataset
Stage
Downloads
rh-clean-control-sft 0 0
Clean Control SFT Mixture A clean SFT mixture dataset for use as a control in reward hacking experiments. This dataset contains only beni...
451
fake-cifarnet 0 0
This is a dataset of "fake" CIFARNet images which were sampled from a high-entropy distribution whose mean and covariance matrix matches tha...
582
fake-cifar10 0 0
This is a dataset of "fake" CIFAR-10 images which were sampled from a high-entropy distribution whose mean and covariance matrix matches tha...
651
fake-mnist 0 0
This is a dataset of "fake" MNIST images which were sampled from a high-entropy distribution whose mean and covariance matrix matches that o...
747
Dataset
Stage
Downloads
fake-mnist 0 0
This is a dataset of "fake" MNIST images which were sampled from a high-entropy distribution whose mean and covariance matrix matches that o...
747
fake-cifar10 0 0
This is a dataset of "fake" CIFAR-10 images which were sampled from a high-entropy distribution whose mean and covariance matrix matches tha...
651
fake-cifarnet 0 0
This is a dataset of "fake" CIFARNet images which were sampled from a high-entropy distribution whose mean and covariance matrix matches tha...
582
rh-clean-control-sft 0 0
Clean Control SFT Mixture A clean SFT mixture dataset for use as a control in reward hacking experiments. This dataset contains only beni...
451
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration