The Decoder· Jonathan Kemper·· 1 day agoAI Score59
Taste-Bench tests whether AI agents make good long-term decisions, and top models barely beat chance
AI agents have bad taste in long-term decisions and a new benchmark proves it
AI summary
Researchers introduced Taste-Bench, a 502-question benchmark that freezes agent trajectories at decision forks and asks models to pick the better of two paths. The best model.
Source: The Decoder · the-decoder.com