Thanks for following through.
If anyone wants to make a proper harness in the future, I think probably the most interesting question here is if the LLM can learn from multiple playthroughs, unlocking harder difficulties, etc.
Modern LLMs, maybe through notetaking?
Thanks for following through.
If anyone wants to make a proper harness in the future, I think probably the most interesting question here is if the LLM can learn from multiple playthroughs, unlocking harder difficulties, etc.
Modern LLMs, maybe through notetaking?