We view Large Language Models as stochastic language layers in a network, where the learnable parameters are the natural language prompts at each layer. We stack two such layers, feeding the output of one layer to the next. We call the stacked architecture a Deep Language Network - DLN
The idea is with a global LLM Registry, it will be easier to get the total_cost (in number of tokens) used by the different LLMs during training and at test.
The idea is with a global LLM Registry, it will be easier to get the total_cost (in number of tokens) used by the different LLMs during training and at test.