Get the app
← back to catalog
AUTHOR

giskardai

No models indexed under this account. Publishes 4 alignment datasets (12.5k downloads) tracked in our /datasets catalog.

Alignment datasets by giskardai

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
phare 0 0
Phare Benchmark Phare is a multilingual benchmark that measures LLM Safety across multiple categories of vulnerabilities, including hallu...
extract
7.5k
realharm 0 0
RealHarm RealHarm is a collection of harmful real-world interactions with AI agents. Dataset Details Dataset ...
misc
2.8k
harmbench-scenarios 0 0
HarmBench Scenarios Safety-evaluation scenarios derived from the HarmBench behavior dataset, serialized as giskard Scenario objects (one ...
evaluate
1.2k
do-not-answer-scenarios 0 0
Do-Not-Answer Scenarios Refusal-evaluation scenarios derived from the Do-Not-Answer dataset, serialized as giskard Scenario objects (one ...
evaluate
1k
Dataset
Stage
Downloads
harmbench-scenarios 0 0
HarmBench Scenarios Safety-evaluation scenarios derived from the HarmBench behavior dataset, serialized as giskard Scenario objects (one ...
evaluate
1.2k
do-not-answer-scenarios 0 0
Do-Not-Answer Scenarios Refusal-evaluation scenarios derived from the Do-Not-Answer dataset, serialized as giskard Scenario objects (one ...
evaluate
1k
realharm 0 0
RealHarm RealHarm is a collection of harmful real-world interactions with AI agents. Dataset Details Dataset ...
misc
2.8k
phare 0 0
Phare Benchmark Phare is a multilingual benchmark that measures LLM Safety across multiple categories of vulnerabilities, including hallu...
extract
7.5k
Dataset
Stage
Downloads
phare 0 0
Phare Benchmark Phare is a multilingual benchmark that measures LLM Safety across multiple categories of vulnerabilities, including hallu...
extract
7.5k
realharm 0 0
RealHarm RealHarm is a collection of harmful real-world interactions with AI agents. Dataset Details Dataset ...
misc
2.8k
harmbench-scenarios 0 0
HarmBench Scenarios Safety-evaluation scenarios derived from the HarmBench behavior dataset, serialized as giskard Scenario objects (one ...
evaluate
1.2k
do-not-answer-scenarios 0 0
Do-Not-Answer Scenarios Refusal-evaluation scenarios derived from the Do-Not-Answer dataset, serialized as giskard Scenario objects (one ...
evaluate
1k
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.

Get the app Browse the catalog
Open in Abliteration