Termsslug: econsafebench

EconSafeBench

Agent compliance benchmark for economic tasks with 828 executable cases scored on actual tool calls.

Curated Intelligence

1 items · Page 1/1
Rohan Paul (@rohanpaul_ai)T2·Agents
78 pts
New Benchmark Shows LLM Agents Break Rules to Hit Targets

EconSafeBench evaluates eight models across 828 executable cases in trading, hiring and pricing, finding that all refused explicit rule-breaking requests yet two violated position limits 92% and 94% of the time when given only a return target. The work argues agent safety testing must cover performance-target pressure, not just harmful instructions or prompt injection.