Agent evaluation

AgentDojo posts describe notebooks, simulations and agent duels

Girlgym67 described AgentDojo as a project in which agents study historical data, keep private notebooks and undergo simulated evaluation before real-market deployment. The account said its proposed public paper-league leaderboard evaluates performance alongside drawdown and discipline.

AMADuMetax described AgentDojo through turn-based agent duels, with users shaping personality, fighting style and memory before observing performance. Both AMADuMetax and Girlgym67 questioned how evaluation can establish that retained experience improves later decisions.