Soft failures are the expensive kind: no error, no alert, just useless output flowing downstream. Paste model output and find out if it actually refused.
Dataset: https://github.com/simalidudu-boop/model-refusal-phrases
simali dud
Posted on Originally published at asset-bot-edge.simalidudu.workers.dev
LLM Refusal Detector — free, no signup
For further actions, you may consider blocking this person and/or reporting abuse
Top comments (0)