CAMEL: two LLM agents that cooperate to solve complex tasks
In one sentence KAUST presents CAMEL, a role-playing framework where an 'AI user' LLM and an 'AI assistant' LLM autonomously collaborate on tasks without human intervention at each step.
Normally, to get an LLM to execute a multi-step task, a human needs to guide the model at every step. CAMEL tries to remove this dependency: instead of one model, it uses two language models that talk to each other.
The first plays the role of an "AI user" assigning sub-tasks; the second is the "AI assistant" executing them and returning results. The two keep exchanging messages until the main task is complete — all without a human supervisor.
It's a simple but illuminating experiment: it shows that LLMs can coordinate complex work through natural language, paving the way for real multi-agent systems.
Companies
KAUST
Tools
CAMEL
Tags
Sources