llama-guard
llama-guard-safety-eval
Associated Paper
Synthetic Multi-Label Safety Dataset for LLaMA Guard 2 & 3
Dataset Summary
This dataset is a synthetic, multi-label safety evaluation corpus designed to align with the LLaMA Guard 2 and LLaMA Guard 3 taxonomies and formats.
Because LLaMA Guard provides no official test datasets or public benchmark aligned with its taxonomy, we construct a fully synthetic evaluation set using a controlled multi-step process. Each example:
Violates 2–3 distinct… See the full description on the dataset page: https://huggingface.co/datasets/AuroraQuantum/llama-guard-safety-eval.details_guardrail__llama-2-7b-guanaco-instruct-sharded
Dataset Card for Evaluation run of guardrail/llama-2-7b-guanaco-instruct-sharded
Dataset Summary
Dataset automatically created during the evaluation run of model guardrail/llama-2-7b-guanaco-instruct-sharded on the Open LLM Leaderboard.
The dataset is composed of 64 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_guardrail__llama-2-7b-guanaco-instruct-sharded.ThinkSafe-4B-n4-filtered-LlamaGuardThinkSafe-0.6B-n4-filtered-LlamaGuardThinkSafe-8B-n4-filtered-LlamaGuardThinkSafe-1.7B-n4-filtered-LlamaGuard