Alignment datasets by RumiaChannel

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.

3 in /datasets
Dataset
Stage
Downloads
harmful_behaviors_ja_synth 0 0
harmful_behaviors_ja_synth Japanese synthetic harmful-behavior prompts for safety/refusal evaluation. Columns: text: prompt text
279
harmless_behaviors_ja_synth 0 0
harmless_behaviors_ja_synth Japanese synthetic harmless instruction prompts for ordinary-response / refusal-direction evaluation. ...
193
harmful_behaviors_ja 0 0
harmful_behaviors_ja_synth mlabonne/harmful_behaviors を DeepSeek v4 pro を用いて日本語訳したものです Japanese synthetic harmful-behavior prompts for sa...
176
Dataset
Stage
Downloads
harmless_behaviors_ja_synth 0 0
harmless_behaviors_ja_synth Japanese synthetic harmless instruction prompts for ordinary-response / refusal-direction evaluation. ...
193
harmful_behaviors_ja_synth 0 0
harmful_behaviors_ja_synth Japanese synthetic harmful-behavior prompts for safety/refusal evaluation. Columns: text: prompt text
279
harmful_behaviors_ja 0 0
harmful_behaviors_ja_synth mlabonne/harmful_behaviors を DeepSeek v4 pro を用いて日本語訳したものです Japanese synthetic harmful-behavior prompts for sa...
176
Dataset
Stage
Downloads
harmful_behaviors_ja_synth 0 0
harmful_behaviors_ja_synth Japanese synthetic harmful-behavior prompts for safety/refusal evaluation. Columns: text: prompt text
279
harmless_behaviors_ja_synth 0 0
harmless_behaviors_ja_synth Japanese synthetic harmless instruction prompts for ordinary-response / refusal-direction evaluation. ...
193
harmful_behaviors_ja 0 0
harmful_behaviors_ja_synth mlabonne/harmful_behaviors を DeepSeek v4 pro を用いて日本語訳したものです Japanese synthetic harmful-behavior prompts for sa...
176
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