Frontier Line and access · Benchmarks evals
Ai2 Releases FairStress LLM Fairness Benchmark
Ai2’s FairStress benchmark finds demographic differences in LLM justifications even when final decisions appear fair.
Read the original at huggingface.coOpens the publisher's site in a new tabAlso covering this
FairStress: A deeper evaluation of fairness in LLM decision-making | Ai2allenai.org, Oct 20FairStress: Ai2 Benchmark Finds LLM Over-Correction Bias | explainx.ai Blog | explainx.aiexplainx.ai, Oct 20FairStress de Ai2: 300 escenarios para medir sesgo en IA – El Ecosistema Startupecosistemastartup.com, Oct 20