Alignment datasets by nayohan

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.

5 in /datasets
Dataset
Stage
Downloads
SQuARe-question 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023square, title={SQuARe: A Large-Scale D...
955
KoSBi-v2 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023kosbi, title={KoSBi: A Dataset for Mit...
772
SQuARe-response 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023square, title={SQuARe: A Large-Scale D...
764
KoSBi 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023kosbi, title={KoSBi: A Dataset for Mit...
750
hh-rlhf-ko 0 0
Translated Anthropic/hh-rlhf using nayohan/llama3-instrucTrans-enko-8b. This dataset is a raw translated dataset and contains repetitive sen...
571
Dataset
Stage
Downloads
hh-rlhf-ko 0 0
Translated Anthropic/hh-rlhf using nayohan/llama3-instrucTrans-enko-8b. This dataset is a raw translated dataset and contains repetitive sen...
571
SQuARe-response 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023square, title={SQuARe: A Large-Scale D...
764
SQuARe-question 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023square, title={SQuARe: A Large-Scale D...
955
KoSBi 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023kosbi, title={KoSBi: A Dataset for Mit...
750
KoSBi-v2 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023kosbi, title={KoSBi: A Dataset for Mit...
772
Dataset
Stage
Downloads
SQuARe-question 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023square, title={SQuARe: A Large-Scale D...
955
KoSBi-v2 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023kosbi, title={KoSBi: A Dataset for Mit...
772
KoSBi 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023kosbi, title={KoSBi: A Dataset for Mit...
750
SQuARe-response 0 0
reference: https://github.com/naver-ai/korean-safety-benchmarks @inproceedings{lee2023square, title={SQuARe: A Large-Scale D...
764
hh-rlhf-ko 0 0
Translated Anthropic/hh-rlhf using nayohan/llama3-instrucTrans-enko-8b. This dataset is a raw translated dataset and contains repetitive sen...
571
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