We have a standard harness for each of the model's that we test. Each prompt includes the rules, access to memory, and a lookup of the complete ruleset. The prompt adapts adding legal actions per turn and guidance depending on the stage of the game (updated based on the technological progress of the player).
Unlike RL algorithms these LLMs wouldn't learn quick enough without the prior knowledge the harness provides
Comments
We have a standard harness for each of the model's that we test. Each prompt includes the rules, access to memory, and a lookup of the complete ruleset. The prompt adapts adding legal actions per turn and guidance depending on the stage of the game (updated based on the technological progress of the player).
Unlike RL algorithms these LLMs wouldn't learn quick enough without the prior knowledge the harness provides
what do you use for memory?
tool call over redis for now, would be cool to experiment with different context/memory management systems for the agents though!