FLEX (Fairness Benchmark in LLM under Extreme Scenarios) - OECD. AI
FLEX is a benchmark for testing whether large language models stay fair when prompts are designed to induce bias . It was developed by researchers ...
Read the original at OECD.AI ↗
- Topics
- Bias & FairnessChatbots & LLMs
- Tracked via
- “AI Bias”