Alignment datasets by kasadin

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.

2 in /datasets
Dataset
Stage
Downloads
alignment-4type 0 0
Alignment 4-Type Dataset 8-type alignment data for over-refusal research. 4 prompt categories × 2 response behaviors. Data De...
170
bif-qg-safe-top120-eval-inference 0 0
BIF QG Safe Top120 (gamma=1000, nbeta=1000) - Eval Inference Results Inference results from evaluating the BIF per-query-group (safe) top...
115
Dataset
Stage
Downloads
alignment-4type 0 0
Alignment 4-Type Dataset 8-type alignment data for over-refusal research. 4 prompt categories × 2 response behaviors. Data De...
170
bif-qg-safe-top120-eval-inference 0 0
BIF QG Safe Top120 (gamma=1000, nbeta=1000) - Eval Inference Results Inference results from evaluating the BIF per-query-group (safe) top...
115
Dataset
Stage
Downloads
bif-qg-safe-top120-eval-inference 0 0
BIF QG Safe Top120 (gamma=1000, nbeta=1000) - Eval Inference Results Inference results from evaluating the BIF per-query-group (safe) top...
115
alignment-4type 0 0
Alignment 4-Type Dataset 8-type alignment data for over-refusal research. 4 prompt categories × 2 response behaviors. Data De...
170
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