HarvestBench Finds Models Pay Unequally to Avoid Killing
An agent can sound aligned and still drive over an animal when the route gets expensive. HarvestBench puts that tradeoff inside a reproducible farm simulation: across 7,201 priced decisions and nine models, kill rates ranged from 0.4% to 98.8%, independent of capability. A short morality briefing pushed five of six reasoning models below 6% kills, while removing it sent all six above 84%, exposing how much behavior depends on framing rather than durable preference.