CircleGuardBench: likes over time

whitecircle/circle-guard-bench is a gradio Space on the Hugging Face Hub. It has 32 likes, 0 in the last 7 days and 2 in the last 30, and ranks #3,941 among Spaces by likes.

First benchmark testing LLM guards on safety and accuracy.

What it uses

• Qwen/Qwen3-0.6B (model)
• Qwen/Qwen3-8B (model)
• Qwen/Qwen2.5-7B-Instruct (model)
• Qwen/Qwen3-4B (model)
• meta-llama/Llama-3.1-8B-Instruct (model)
• Qwen/Qwen3-1.7B (model)
• meta-llama/Llama-3.2-3B-Instruct (model)
• google/gemma-3-4b-it (model)
• mistralai/Mistral-Small-3.1-24B-Instruct-2503 (model)
• google/gemma-3-12b-it (model)
• meta-llama/Llama-4-Scout-17B-16E-Instruct (model)
• mistralai/Ministral-8B-Instruct-2410 (model)
• meta-llama/Llama-Guard-4-12B (model)
• meta-llama/Llama-Guard-3-8B (model)
• meta-llama/Llama-Guard-3-1B (model)
• meta-llama/Llama-4-Maverick-17B-128E-Instruct (model)
• AtlaAI/Selene-1-Mini-Llama-3.1-8B (model)

Open whitecircle/circle-guard-bench on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.